{ | |
"_attn_implementation_autoset": true, | |
"_name_or_path": "hf-internal-testing/tiny-random-GPTNeoForSequenceClassification", | |
"activation_function": "gelu_new", | |
"architectures": [ | |
"GPTNeoForSequenceClassification" | |
], | |
"attention_dropout": 0.0, | |
"attention_layers": [ | |
"global", | |
"local", | |
"global", | |
"local" | |
], | |
"attention_types": [ | |
[ | |
[ | |
"global", | |
"local" | |
], | |
2 | |
] | |
], | |
"bos_token_id": 0, | |
"classifier_dropout": 0.1, | |
"embed_dropout": 0.0, | |
"eos_token_id": 0, | |
"hidden_size": 32, | |
"initializer_range": 0.02, | |
"intermediate_size": null, | |
"layer_norm_epsilon": 1e-05, | |
"max_position_embeddings": 512, | |
"model_type": "gpt_neo", | |
"num_heads": 4, | |
"num_layers": 4, | |
"pad_token_id": 1023, | |
"resid_dropout": 0.0, | |
"summary_activation": null, | |
"summary_first_dropout": 0.1, | |
"summary_proj_to_labels": true, | |
"summary_type": "cls_index", | |
"summary_use_proj": true, | |
"transformers_version": "4.48.2", | |
"use_cache": true, | |
"vocab_size": 1024, | |
"window_size": 7 | |
} | |