Baichuan-13B-Chat-sft-super / finetuning_args.json
wangrongsheng's picture
update model
6da17dd
{
"finetuning_type": "lora",
"lora_alpha": 32.0,
"lora_dropout": 0.1,
"lora_rank": 8,
"lora_target": [
"W_pack"
],
"name_module_trainable": "mlp",
"num_hidden_layers": 32,
"num_layer_trainable": 3
}