Model: QpiEImitation/gkd_gsm8k_S-Qwen2-1.5B-Instruct_T-Qwen2-7B-Instruct Source: Original Platform
20 lines
500 B
JSON
20 lines
500 B
JSON
{
|
|
"method": "gkd",
|
|
"student_model": "Qwen/Qwen2-1.5B-Instruct",
|
|
"teacher_model": "Qwen/Qwen2-7B-Instruct",
|
|
"dataset": "gsm8k/main",
|
|
"train_samples": 2000,
|
|
"eval_samples": 256,
|
|
"beta": 0.5,
|
|
"max_steps": 800,
|
|
"num_epochs": 1.0,
|
|
"lr": 2e-06,
|
|
"batch_size": 1,
|
|
"grad_accum": 8,
|
|
"output_dir": "gkd_gsm8k_S-Qwen2-1.5B-Instruct_T-Qwen2-7B-Instruct_20260420_154424",
|
|
"use_lora": false,
|
|
"lora_r": null,
|
|
"lora_alpha": null,
|
|
"lora_dropout": null,
|
|
"lora_target_modules": null
|
|
} |