zannvznn/qwen3-0.6b-math-l45-qlora-merged-fp16-v2
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: zannvznn/qwen3-0.6b-math-l45-qlora-merged-fp16-v2
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-08-05 22:57:17 +08:00
commit 5893f5402f
11 changed files with 350 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
run_config.json Normal file
View File

@@ -0,0 +1,24 @@
{
"model_id": "Qwen/Qwen3-0.6B-Base",
"run_name": "qwen3-0.6b-math-l45-qlora-sft-v2",
"hf_repo_id": "zannvznn/qwen3-0.6b-math-l45-qlora-merged-fp16-v2",
"train_file": "/kaggle/input/datasets/anurhalizah/math-he/math_level45_train.parquet",
"test_file": "/kaggle/input/datasets/anurhalizah/math-he/math_level45_test.parquet",
"train_subset_size": 3994,
"train_rows_used": 3994,
"loss_eval_rows_used": 200,
"max_seq_length": 2048,
"num_train_epochs": 3,
"learning_rate": 0.0002,
"per_device_train_batch_size": 1,
"per_device_eval_batch_size": 1,
"gradient_accumulation_steps": 8,
"fp16": true,
"bf16": false,
"qlora": true,
"shuffle_train": true,
"world_size": 1,
"device_map": "balanced",
"max_memory": "{0: 13958643712, 1: 13958643712}",
"final_eval_note": "Final 500-question evaluation is intentionally separate."
}