pretrain_w-cot_w-asd / adapter_config.json
slplab's picture
Training in progress, step 2000
372bd56 verified
raw
history blame contribute delete
No virus
4.95 kB
{
"alpha_pattern": {},
"auto_mapping": null,
"base_model_name_or_path": "EleutherAI/polyglot-ko-1.3b",
"bias": "none",
"fan_in_fan_out": false,
"inference_mode": true,
"init_lora_weights": true,
"layer_replication": null,
"layers_pattern": null,
"layers_to_transform": null,
"loftq_config": {},
"lora_alpha": 16,
"lora_dropout": 0.05,
"megatron_config": null,
"megatron_core": "megatron.core",
"modules_to_save": null,
"peft_type": "LORA",
"r": 8,
"rank_pattern": {},
"revision": null,
"target_modules": [
"gpt_neox.layers.23.mlp.dense_4h_to_h",
"gpt_neox.layers.20.attention.query_key_value",
"gpt_neox.layers.4.mlp.dense_h_to_4h",
"gpt_neox.layers.5.mlp.dense_h_to_4h",
"gpt_neox.layers.11.mlp.dense_h_to_4h",
"gpt_neox.layers.20.mlp.dense_h_to_4h",
"gpt_neox.layers.3.attention.query_key_value",
"gpt_neox.layers.13.mlp.dense_4h_to_h",
"gpt_neox.layers.12.mlp.dense_h_to_4h",
"gpt_neox.layers.17.mlp.dense_h_to_4h",
"gpt_neox.layers.22.attention.query_key_value",
"gpt_neox.layers.2.attention.dense",
"gpt_neox.layers.15.mlp.dense_4h_to_h",
"gpt_neox.layers.2.mlp.dense_h_to_4h",
"gpt_neox.layers.7.attention.dense",
"gpt_neox.layers.8.mlp.dense_h_to_4h",
"gpt_neox.layers.16.mlp.dense_4h_to_h",
"gpt_neox.layers.15.attention.dense",
"gpt_neox.layers.9.mlp.dense_h_to_4h",
"gpt_neox.layers.22.attention.dense",
"gpt_neox.layers.20.mlp.dense_4h_to_h",
"gpt_neox.layers.0.attention.query_key_value",
"gpt_neox.layers.8.attention.dense",
"gpt_neox.layers.14.mlp.dense_h_to_4h",
"gpt_neox.layers.1.attention.dense",
"gpt_neox.layers.6.mlp.dense_h_to_4h",
"gpt_neox.layers.1.mlp.dense_h_to_4h",
"gpt_neox.layers.13.mlp.dense_h_to_4h",
"gpt_neox.layers.17.attention.dense",
"gpt_neox.layers.19.mlp.dense_h_to_4h",
"gpt_neox.layers.6.attention.dense",
"gpt_neox.layers.4.mlp.dense_4h_to_h",
"gpt_neox.layers.8.mlp.dense_4h_to_h",
"gpt_neox.layers.19.attention.dense",
"gpt_neox.layers.19.mlp.dense_4h_to_h",
"gpt_neox.layers.18.attention.dense",
"gpt_neox.layers.0.attention.dense",
"gpt_neox.layers.15.mlp.dense_h_to_4h",
"gpt_neox.layers.10.attention.query_key_value",
"gpt_neox.layers.18.mlp.dense_h_to_4h",
"gpt_neox.layers.9.attention.dense",
"gpt_neox.layers.23.mlp.dense_h_to_4h",
"gpt_neox.layers.11.attention.query_key_value",
"gpt_neox.layers.5.attention.dense",
"gpt_neox.layers.18.mlp.dense_4h_to_h",
"gpt_neox.layers.23.attention.query_key_value",
"gpt_neox.layers.13.attention.query_key_value",
"gpt_neox.layers.17.attention.query_key_value",
"gpt_neox.layers.4.attention.dense",
"gpt_neox.layers.8.attention.query_key_value",
"gpt_neox.layers.17.mlp.dense_4h_to_h",
"gpt_neox.layers.20.attention.dense",
"gpt_neox.layers.0.mlp.dense_4h_to_h",
"gpt_neox.layers.5.mlp.dense_4h_to_h",
"gpt_neox.layers.0.mlp.dense_h_to_4h",
"gpt_neox.layers.3.attention.dense",
"gpt_neox.layers.2.attention.query_key_value",
"gpt_neox.layers.21.attention.dense",
"gpt_neox.layers.7.mlp.dense_h_to_4h",
"gpt_neox.layers.5.attention.query_key_value",
"gpt_neox.layers.18.attention.query_key_value",
"gpt_neox.layers.6.mlp.dense_4h_to_h",
"gpt_neox.layers.15.attention.query_key_value",
"gpt_neox.layers.22.mlp.dense_4h_to_h",
"gpt_neox.layers.6.attention.query_key_value",
"gpt_neox.layers.10.mlp.dense_h_to_4h",
"gpt_neox.layers.14.attention.query_key_value",
"gpt_neox.layers.10.mlp.dense_4h_to_h",
"gpt_neox.layers.1.attention.query_key_value",
"gpt_neox.layers.11.attention.dense",
"gpt_neox.layers.14.mlp.dense_4h_to_h",
"gpt_neox.layers.16.attention.dense",
"gpt_neox.layers.21.attention.query_key_value",
"gpt_neox.layers.16.attention.query_key_value",
"gpt_neox.layers.4.attention.query_key_value",
"gpt_neox.layers.7.attention.query_key_value",
"gpt_neox.layers.1.mlp.dense_4h_to_h",
"gpt_neox.layers.16.mlp.dense_h_to_4h",
"gpt_neox.layers.21.mlp.dense_h_to_4h",
"gpt_neox.layers.2.mlp.dense_4h_to_h",
"gpt_neox.layers.9.mlp.dense_4h_to_h",
"gpt_neox.layers.22.mlp.dense_h_to_4h",
"gpt_neox.layers.11.mlp.dense_4h_to_h",
"gpt_neox.layers.7.mlp.dense_4h_to_h",
"gpt_neox.layers.23.attention.dense",
"gpt_neox.layers.12.attention.dense",
"gpt_neox.layers.13.attention.dense",
"gpt_neox.layers.3.mlp.dense_h_to_4h",
"gpt_neox.layers.3.mlp.dense_4h_to_h",
"gpt_neox.layers.12.attention.query_key_value",
"gpt_neox.layers.12.mlp.dense_4h_to_h",
"gpt_neox.layers.19.attention.query_key_value",
"gpt_neox.layers.21.mlp.dense_4h_to_h",
"gpt_neox.layers.10.attention.dense",
"gpt_neox.layers.14.attention.dense",
"gpt_neox.layers.9.attention.query_key_value"
],
"task_type": "CAUSAL_LM",
"use_dora": false,
"use_rslora": false
}