21 lines
727 B
JSON
21 lines
727 B
JSON
{
|
|
"base_model": "/home/edgeai_lab/minseong_results/rsn_tuned_qwen2_5_7b-instruct_lr5e-5_20260715_124936",
|
|
"fine_tuning_type": "GSM8K Fine-tuning with Safety Neuron Freezing",
|
|
"safety_neurons_file": "/home/edgeai_lab/minseong_results/critical_safety_qwen2_5_7b-instruct/critical_safety_neurons.txt",
|
|
"dataset": "GSM8K",
|
|
"num_train_samples": 7473,
|
|
"batch_size": 4,
|
|
"grad_accum": 4,
|
|
"learning_rate": 5e-05,
|
|
"weight_decay": 0.01,
|
|
"warmup_ratio": 0.1,
|
|
"epochs": 3,
|
|
"max_length": 1024,
|
|
"max_grad_norm": 1.0,
|
|
"lr_scheduler_type": "cosine",
|
|
"optimizer": "adamw_torch",
|
|
"gradient_checkpointing": false,
|
|
"dtype": "bf16",
|
|
"trainer_type": "Trainer",
|
|
"strategy": "Freeze safety neurons, train others"
|
|
} |