|
{
|
|
"_name_": "Judge-GPT2",
|
|
"_name_or_path": "Wonder-Griffin/JudgeLLM2",
|
|
"activation_function": "gelu_new",
|
|
"architectures": [
|
|
"GPT2LMHeadModel"
|
|
],
|
|
"attn_pdrop": 0.1,
|
|
"batch_size": 32,
|
|
"bias": true,
|
|
"block_size": 512,
|
|
"bos_token_id": 50256,
|
|
"dim_feedforward": 3072,
|
|
"dropout": 0.1,
|
|
"embd_pdrop": 0.1,
|
|
"eos_token_id": 50256,
|
|
"ff_expansion_factor": 4,
|
|
"hidden_act": "gelu",
|
|
"id2label": {
|
|
"0": "LABEL_0",
|
|
"1": "LABEL_1",
|
|
"2": "LABEL_2",
|
|
"3": "LABEL_3",
|
|
"4": "LABEL_4"
|
|
},
|
|
"inference_mode": true,
|
|
"initializer_range": 0.02,
|
|
"label2id": {
|
|
"LABEL_0": 0,
|
|
"LABEL_1": 1,
|
|
"LABEL_2": 2,
|
|
"LABEL_3": 3,
|
|
"LABEL_4": 4
|
|
},
|
|
"label_smoothing": 0.1,
|
|
"layer_norm_epsilon": 1e-05,
|
|
"learning_rate": 0.0003,
|
|
"log_interval": 100,
|
|
"max_grad_norm": 1.0,
|
|
"model_type": "gpt2",
|
|
"n_embd": 768,
|
|
"n_head": 12,
|
|
"n_inner": null,
|
|
"n_layer": 12,
|
|
"n_positions": 512,
|
|
"output_dir": "C:/Users/wonde/output",
|
|
"pretrained_weights": "Wonder-Griffin/Judge-GPT2",
|
|
"reorder_and_upcast_attn": false,
|
|
"resid_pdrop": 0.1,
|
|
"scale_attn_by_inverse_layer_idx": false,
|
|
"scale_attn_weights": true,
|
|
"summary_activation": null,
|
|
"summary_first_dropout": 0.1,
|
|
"summary_proj_to_labels": true,
|
|
"summary_type": "cls_index",
|
|
"summary_use_proj": true,
|
|
"task_heads": {
|
|
"classifier_head": {
|
|
"params": {
|
|
"num_labels": 5
|
|
},
|
|
"type": "JudgeClassifier"
|
|
},
|
|
"lm_head": {
|
|
"params": {
|
|
"vocab_size": 50257
|
|
},
|
|
"type": "JudgeCasualLMHead"
|
|
},
|
|
"qa_head": {
|
|
"params": {
|
|
"num_labels": 2
|
|
},
|
|
"type": "JudgeWithQA"
|
|
}
|
|
},
|
|
"task_specific_params": {
|
|
"question-answering": {
|
|
"max_answer_length": 100
|
|
},
|
|
"sequence-classification": {
|
|
"eval_steps": 500
|
|
},
|
|
"text-generation": {
|
|
"do_sample": true,
|
|
"max_length": 100
|
|
}
|
|
},
|
|
"tokenizer": {
|
|
"params": {
|
|
"vocab_size": 50257
|
|
},
|
|
"type": "AutoTokenizer"
|
|
},
|
|
"torch_dtype": "float32",
|
|
"total_steps": 10000,
|
|
"transformers_version": "4.43.3",
|
|
"use_cache": true,
|
|
"vocab_size": 30522,
|
|
"warmup_steps": 1000,
|
|
"weight_decay": 0.01
|
|
}
|
|
|