|
{ |
|
"_name_or_path": "EleutherAI/pythia-160m", |
|
"architectures": [ |
|
"GPTNeoXForSequenceClassification" |
|
], |
|
"attention_bias": true, |
|
"attention_dropout": 0.0, |
|
"bos_token_id": 0, |
|
"classifier_dropout": 0.1, |
|
"eos_token_id": 0, |
|
"finetuning_task": "text-classification", |
|
"hidden_act": "gelu", |
|
"hidden_dropout": 0.0, |
|
"hidden_size": 768, |
|
"id2label": { |
|
"0": "a", |
|
"1": "b", |
|
"2": "c", |
|
"3": "d", |
|
"4": "e" |
|
}, |
|
"initializer_range": 0.02, |
|
"intermediate_size": 3072, |
|
"label2id": { |
|
"a": 0, |
|
"b": 1, |
|
"c": 2, |
|
"d": 3, |
|
"e": 4 |
|
}, |
|
"layer_norm_eps": 1e-05, |
|
"max_position_embeddings": 2048, |
|
"model_type": "gpt_neox", |
|
"num_attention_heads": 12, |
|
"num_hidden_layers": 12, |
|
"pad_token_id": 0, |
|
"problem_type": "single_label_classification", |
|
"rope_scaling": null, |
|
"rotary_emb_base": 10000, |
|
"rotary_pct": 0.25, |
|
"tie_word_embeddings": false, |
|
"torch_dtype": "float32", |
|
"transformers_version": "4.43.2", |
|
"use_cache": true, |
|
"use_parallel_residual": true, |
|
"vocab_size": 50304 |
|
} |
|
|