Files
gkd_math500_S-Qwen2.5-3B-In…/run_config.json
ModelHub XC 9600347571 初始化项目,由ModelHub XC社区提供模型
Model: QpiEImitation/gkd_math500_S-Qwen2.5-3B-Instruct_T-Qwen2-7B-Instruct
Source: Original Platform
2026-06-16 03:43:16 +08:00

21 lines
571 B
JSON

{
"method": "gkd",
"student_model": "Qwen/Qwen2.5-3B-Instruct",
"teacher_model": "Qwen/Qwen2-7B-Instruct",
"train_dataset": "EleutherAI/hendrycks_math/all",
"eval_dataset": "HuggingFaceH4/MATH-500",
"train_samples": 7500,
"eval_samples": 500,
"beta": 0.5,
"max_steps": 800,
"num_epochs": 1.0,
"lr": 2e-06,
"batch_size": 1,
"grad_accum": 8,
"output_dir": "gkd_math500_S-Qwen2.5-3B-Instruct_T-Qwen2-7B-Instruct_20260420_154438",
"use_lora": false,
"lora_r": null,
"lora_alpha": null,
"lora_dropout": null,
"lora_target_modules": null
}