platistil / adapter_config.json
sirr
End of training
dc765e3
{
"auto_mapping": null,
"base_model_name_or_path": "gpt2-medium",
"bias": "none",
"fan_in_fan_out": false,
"inference_mode": true,
"init_lora_weights": true,
"layers_pattern": null,
"layers_to_transform": null,
"lora_alpha": 32,
"lora_dropout": 0.1,
"modules_to_save": null,
"peft_type": "LORA",
"r": 8,
"revision": null,
"target_modules": [
"transformer.h.0.attn.c_attn",
"transformer.h.0.attn.c_proj",
"transformer.h.1.attn.c_attn",
"transformer.h.1.attn.c_proj",
"transformer.h.2.attn.c_attn",
"transformer.h.2.attn.c_proj",
"transformer.h.3.attn.c_attn",
"transformer.h.3.attn.c_proj",
"transformer.h.4.attn.c_attn",
"transformer.h.4.attn.c_proj",
"transformer.h.5.attn.c_attn",
"transformer.h.5.attn.c_proj",
"transformer.h.6.attn.c_attn",
"transformer.h.6.attn.c_proj",
"transformer.h.7.attn.c_attn",
"transformer.h.7.attn.c_proj",
"transformer.h.8.attn.c_attn",
"transformer.h.8.attn.c_proj",
"transformer.h.9.attn.c_attn",
"transformer.h.9.attn.c_proj",
"transformer.h.10.attn.c_attn",
"transformer.h.10.attn.c_proj",
"transformer.h.11.attn.c_attn",
"transformer.h.11.attn.c_proj",
"transformer.h.12.attn.c_attn",
"transformer.h.12.attn.c_proj",
"transformer.h.13.attn.c_attn",
"transformer.h.13.attn.c_proj",
"transformer.h.14.attn.c_attn",
"transformer.h.14.attn.c_proj",
"transformer.h.15.attn.c_attn",
"transformer.h.15.attn.c_proj",
"transformer.h.16.attn.c_attn",
"transformer.h.16.attn.c_proj",
"transformer.h.17.attn.c_attn",
"transformer.h.17.attn.c_proj",
"transformer.h.18.attn.c_attn",
"transformer.h.18.attn.c_proj",
"transformer.h.19.attn.c_attn",
"transformer.h.19.attn.c_proj",
"transformer.h.20.attn.c_attn",
"transformer.h.20.attn.c_proj",
"transformer.h.21.attn.c_attn",
"transformer.h.21.attn.c_proj",
"transformer.h.22.attn.c_attn",
"transformer.h.22.attn.c_proj",
"transformer.h.23.attn.c_attn",
"transformer.h.23.attn.c_proj",
"transformer.h.0.mlp.c_fc",
"transformer.h.0.mlp.c_proj",
"transformer.h.1.mlp.c_fc",
"transformer.h.1.mlp.c_proj",
"transformer.h.2.mlp.c_fc",
"transformer.h.2.mlp.c_proj",
"transformer.h.3.mlp.c_fc",
"transformer.h.3.mlp.c_proj",
"transformer.h.4.mlp.c_fc",
"transformer.h.4.mlp.c_proj",
"transformer.h.5.mlp.c_fc",
"transformer.h.5.mlp.c_proj",
"transformer.h.6.mlp.c_fc",
"transformer.h.6.mlp.c_proj",
"transformer.h.7.mlp.c_fc",
"transformer.h.7.mlp.c_proj",
"transformer.h.8.mlp.c_fc",
"transformer.h.8.mlp.c_proj",
"transformer.h.9.mlp.c_fc",
"transformer.h.9.mlp.c_proj",
"transformer.h.10.mlp.c_fc",
"transformer.h.10.mlp.c_proj",
"transformer.h.11.mlp.c_fc",
"transformer.h.11.mlp.c_proj",
"transformer.h.12.mlp.c_fc",
"transformer.h.12.mlp.c_proj",
"transformer.h.13.mlp.c_fc",
"transformer.h.13.mlp.c_proj",
"transformer.h.14.mlp.c_fc",
"transformer.h.14.mlp.c_proj",
"transformer.h.15.mlp.c_fc",
"transformer.h.15.mlp.c_proj",
"transformer.h.16.mlp.c_fc",
"transformer.h.16.mlp.c_proj",
"transformer.h.17.mlp.c_fc",
"transformer.h.17.mlp.c_proj",
"transformer.h.18.mlp.c_fc",
"transformer.h.18.mlp.c_proj",
"transformer.h.19.mlp.c_fc",
"transformer.h.19.mlp.c_proj",
"transformer.h.20.mlp.c_fc",
"transformer.h.20.mlp.c_proj",
"transformer.h.21.mlp.c_fc",
"transformer.h.21.mlp.c_proj",
"transformer.h.22.mlp.c_fc",
"transformer.h.22.mlp.c_proj",
"transformer.h.23.mlp.c_fc",
"transformer.h.23.mlp.c_proj"
],
"task_type": "CAUSAL_LM"
}