初始化项目,由ModelHub XC社区提供模型
Model: Tencent-Hunyuan/Hy-MT2-7B-GGUF Source: Original Platform
This commit is contained in:
6
train/deepspeed_support/merge_lora_weight.sh
Normal file
6
train/deepspeed_support/merge_lora_weight.sh
Normal file
@@ -0,0 +1,6 @@
|
||||
SCRIPT_DIR=$(dirname "$0")
|
||||
python3 ${SCRIPT_DIR}/merge_lora_weight.py \
|
||||
--base_model_path /xxx/hy_l_train/checkpoint-200 \
|
||||
--adapter_model_path /xxx/runs/hy_l_lora_train/checkpoint-200 \
|
||||
--output_path /xxx/ckpts/merged_hy_lora_weight \
|
||||
--save_dtype bf16
|
||||
Reference in New Issue
Block a user