./slurm_setup.sh: line 48: GLOBAL_TRAIN_BATCH_SIZE / NNODES / GPUS_PER_NODE / GRADIENT_ACCUMULATION_STEPS: division by 0 (error token is "GRADIENT_ACCUMULATION_STEPS") ./slurm_setup.sh: line 48: GLOBAL_TRAIN_BATCH_SIZE / NNODES / GPUS_PER_NODE / GRADIENT_ACCUMULATION_STEPS: division by 0 (error token is "GRADIENT_ACCUMULATION_STEPS") ./slurm_setup.sh: line 48: GLOBAL_TRAIN_BATCH_SIZE / NNODES / GPUS_PER_NODE / GRADIENT_ACCUMULATION_STEPS: division by 0 (error token is "GRADIENT_ACCUMULATION_STEPS") ./slurm_setup.sh: line 48: GLOBAL_TRAIN_BATCH_SIZE / NNODES / GPUS_PER_NODE / GRADIENT_ACCUMULATION_STEPS: division by 0 (error token is "GRADIENT_ACCUMULATION_STEPS") ./slurm_setup.sh: line 48: GLOBAL_TRAIN_BATCH_SIZE / NNODES / GPUS_PER_NODE / GRADIENT_ACCUMULATION_STEPS: division by 0 (error token is "GRADIENT_ACCUMULATION_STEPS") ./slurm_setup.sh: line 48: GLOBAL_TRAIN_BATCH_SIZE / NNODES / GPUS_PER_NODE / GRADIENT_ACCUMULATION_STEPS: division by 0 (error token is "GRADIENT_ACCUMULATION_STEPS") ./slurm_setup.sh: line 48: GLOBAL_TRAIN_BATCH_SIZE / NNODES / GPUS_PER_NODE / GRADIENT_ACCUMULATION_STEPS: division by 0 (error token is "GRADIENT_ACCUMULATION_STEPS") ./slurm_setup.sh: line 48: GLOBAL_TRAIN_BATCH_SIZE / NNODES / GPUS_PER_NODE / GRADIENT_ACCUMULATION_STEPS: division by 0 (error token is "GRADIENT_ACCUMULATION_STEPS") 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,855 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,856 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,855 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,856 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,858 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,858 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,858 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,858 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,858 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,858 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,858 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,859 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,859 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,859 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,859 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,859 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,859 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,859 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,859 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,859 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,950 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,950 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,950 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,950 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,950 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,950 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,950 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,950 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,950 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,950 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,950 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,950 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,950 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,950 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,950 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,950 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,965 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,965 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,965 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,965 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,965 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,965 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,965 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,965 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,965 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,965 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,965 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,965 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,965 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,965 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,965 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,965 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,965 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,965 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,965 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,965 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,965 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,965 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,966 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,966 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,971 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,971 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,970 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,970 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,970 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,970 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,970 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,970 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,971 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,971 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,971 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,971 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,971 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,971 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,971 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,971 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,971 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,971 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,971 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,971 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,971 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,971 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,970 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,970 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,984 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,984 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,984 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,984 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,984 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,984 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,984 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,984 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,984 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,984 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,984 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,984 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,985 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,985 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. 2026-04-13 19:56:19,985 - INFO - Note: detected 255 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable. 2026-04-13 19:56:19,985 - INFO - Note: NumExpr detected 255 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8. /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( /home/ligengz/anaconda3/lib/python3.11/site-packages/pandas/core/arrays/masked.py:60: UserWarning: Pandas requires version '1.3.6' or newer of 'bottleneck' (version '1.3.5' currently installed). from pandas.core import ( 2026-04-13 19:56:20,960 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,960 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,961 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,961 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,961 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,961 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,962 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,963 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:20,963 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,055 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,055 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,055 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,055 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,055 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,055 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,055 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,056 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,056 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,056 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,056 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,056 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,056 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,056 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,056 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,057 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,058 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,067 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,067 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,067 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,067 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,067 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,067 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,067 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:21,068 - INFO - PyTorch version 2.6.0 available. 2026-04-13 19:56:58,303 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3339', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:56:58,332 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3578', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:56:58,457 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3035', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 5, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:56:58,458 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3035', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 4, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:56:58,471 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3339', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 4, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:56:58,475 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3339', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 3, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:56:58,491 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3035', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 1, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:56:58,493 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': '2026-04-13 19:56:58,492 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3339', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 5, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3035', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 2, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:56:58,500 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-1062', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 1, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:56:58,500 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-1062', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 4, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:56:58,517 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-1062', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 7, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:56:58,517 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-1062', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 6, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,552 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3035', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,557 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-1062', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,599 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-1062', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 2, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,600 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-1062', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 3, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,603 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3251', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,604 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3035', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 3, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,612 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-1062', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 5, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,613 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3339', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 2, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,615 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3339', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 7, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,618 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3035', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 7, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,622 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3035', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 6, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,632 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3339', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 6, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,637 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3339', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 1, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,684 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-57_batch-block1-1089', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,705 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3578', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 1, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,716 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3578', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 7, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,720 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3578', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 5, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,722 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3578', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 2, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,723 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3578', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 6, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,743 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3578', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 3, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,744 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3578', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 4, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,791 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3251', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 5, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,812 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3251', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 7, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,815 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3251', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 6, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,815 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3251', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 4, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,843 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-57_batch-block1-1089', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 5, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,849 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-57_batch-block1-1089', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 3, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,851 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-57_batch-block1-1089', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 7, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,851 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-57_batch-block1-1089', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 1, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,852 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-57_batch-block1-1089', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 6, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,854 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-57_batch-block1-1089', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 2, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:03,855 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-57_batch-block1-1089', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 4, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:07,900 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3736', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:08,054 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3273', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 6, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:08,054 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3273', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 3, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:08,054 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3273', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 2, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:08,069 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3273', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 7, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:08,070 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3273', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 1, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:08,088 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3736', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 4, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:08,090 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3736', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 6, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:08,090 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3736', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 1, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:08,103 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3736', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 5, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:08,104 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3736', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 7, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:08,107 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3736', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 3, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:08,124 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3736', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 2, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:12,723 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3273', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 0, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:12,762 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3273', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 4, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:12,762 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3273', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 5, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:12,788 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3251', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 2, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:12,790 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3251', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 3, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} 2026-04-13 19:57:12,828 - INFO - Training config: {'model_name': 'Qwen/Qwen3-8B', 'template_name': 'qwen', 'block_size': 40960, 'wandb_project': 'ThreadWeaver', 'train_file_path': './data-04-13', 'dagger': False, 'attn_implementation': 'flex_attention', 'output_dir': 'runs/dev/tw-8node-resume', 'overwrite_output_dir': False, 'do_train': False, 'do_eval': False, 'do_predict': False, 'eval_strategy': , 'prediction_loss_only': False, 'per_device_train_batch_size': 1, 'per_device_eval_batch_size': 1, 'per_gpu_train_batch_size': None, 'per_gpu_eval_batch_size': None, 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': None, 'eval_delay': 0, 'torch_empty_cache_steps': None, 'learning_rate': 1e-05, 'weight_decay': 0.0001, 'adam_beta1': 0.9, 'adam_beta2': 0.95, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 8.0, 'max_steps': -1, 'lr_scheduler_type': , 'lr_scheduler_kwargs': {}, 'warmup_ratio': 0.05, 'warmup_steps': 0, 'log_level': 'passive', 'log_level_replica': 'warning', 'log_on_each_node': True, 'logging_dir': 'runs/dev/tw-8node-resume/runs/Apr13_19-56-56_batch-block1-3251', 'logging_strategy': , 'logging_first_step': False, 'logging_steps': 1.0, 'logging_nan_inf_filter': True, 'save_strategy': , 'save_steps': 20, 'save_total_limit': 2, 'save_safetensors': True, 'save_on_each_node': False, 'save_only_model': False, 'restore_callback_states_from_checkpoint': False, 'no_cuda': False, 'use_cpu': False, 'use_mps_device': False, 'seed': 42, 'data_seed': None, 'jit_mode_eval': False, 'use_ipex': False, 'bf16': True, 'fp16': False, 'fp16_opt_level': 'O1', 'half_precision_backend': 'auto', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': None, 'local_rank': 1, 'ddp_backend': None, 'tpu_num_cores': None, 'tpu_metrics_debug': False, 'debug': [], 'dataloader_drop_last': False, 'eval_steps': None, 'dataloader_num_workers': 0, 'dataloader_prefetch_factor': None, 'past_index': -1, 'run_name': 'runs/dev/tw-8node-resume', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': None, 'load_best_model_at_end': False, 'metric_for_best_model': None, 'greater_is_better': None, 'ignore_data_skip': False, 'fsdp': [], 'fsdp_min_num_params': 0, 'fsdp_config': {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, 'tp_size': 0, 'fsdp_transformer_layer_cls_to_wrap': None, 'accelerator_config': {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False}, 'deepspeed': 'configs/deepspeed_zero3.json', 'label_smoothing_factor': 0.0, 'optim': , 'optim_args': None, 'adafactor': False, 'group_by_length': False, 'length_column_name': 'length', 'report_to': ['wandb'], 'ddp_find_unused_parameters': None, 'ddp_bucket_cap_mb': None, 'ddp_broadcast_buffers': None, 'dataloader_pin_memory': True, 'dataloader_persistent_workers': False, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': False, 'resume_from_checkpoint': None, 'hub_model_id': None, 'hub_strategy': , 'hub_token': None, 'hub_private_repo': None, 'hub_always_push': False, 'gradient_checkpointing': True, 'gradient_checkpointing_kwargs': None, 'include_inputs_for_metrics': False, 'include_for_metrics': [], 'eval_do_concat_batches': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': None, 'push_to_hub_organization': None, 'push_to_hub_token': None, '_n_gpu': 1, 'mp_parameters': '', 'auto_find_batch_size': False, 'full_determinism': False, 'torchdynamo': None, 'ray_scope': 'last', 'ddp_timeout': 1800, 'torch_compile': False, 'torch_compile_backend': None, 'torch_compile_mode': None, 'include_tokens_per_second': False, 'include_num_input_tokens_seen': False, 'neftune_noise_alpha': None, 'optim_target_modules': None, 'batch_eval_metrics': False, 'eval_on_start': False, 'use_liger_kernel': True, 'eval_use_gather_object': False, 'average_tokens_across_devices': True, 'model_init_kwargs': None, 'chat_template_path': None, 'dataset_text_field': 'qwen_text', 'dataset_kwargs': None, 'dataset_num_proc': None, 'eos_token': None, 'pad_token': None, 'max_length': 1024, 'packing': False, 'packing_strategy': 'ffd', 'padding_free': False, 'pad_to_multiple_of': None, 'eval_packing': None, 'completion_only_loss': None, 'assistant_only_loss': False, 'activation_offloading': False, 'max_seq_length': None} Loading checkpoint shards: 0%| | 0/5 [00:00