QpiEImitation/opd_math500_S-Qwen2-0.5B-Instruct_T-Qwen2-7B-Instruct
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: QpiEImitation/opd_math500_S-Qwen2-0.5B-Instruct_T-Qwen2-7B-Instruct
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-06-16 06:06:15 +08:00
commit 4e17d76798
10 changed files with 236 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
run_config.json Normal file
View File

@@ -0,0 +1,28 @@
{
"method": "opd",
"student_model": "Qwen/Qwen2-0.5B-Instruct",
"teacher_model": "Qwen/Qwen2-7B-Instruct",
"train_dataset": "EleutherAI/hendrycks_math/all",
"eval_dataset": "HuggingFaceH4/MATH-500",
"train_samples": 7500,
"eval_samples": 500,
"opd_mode": "expectation",
"trust_region": true,
"ppo_clip_eps": 0.2,
"use_correction": false,
"correction_alpha": 0.2,
"correction_lr": 0.001,
"num_inner_steps": 10,
"replay_buffer_size": 10,
"max_steps": 800,
"num_epochs": 1.0,
"lr": 2e-06,
"batch_size": 1,
"grad_accum": 8,
"output_dir": "opd_math500_S-Qwen2-0.5B-Instruct_T-Qwen2-7B-Instruct_expectation_tr_L10_buf10_20260420_154437",
"use_lora": false,
"lora_r": null,
"lora_alpha": null,
"lora_dropout": null,
"lora_target_modules": null
}