25 lines
865 B
JSON
25 lines
865 B
JSON
{
|
|
"_name_or_path": "codesage/codesage-large-v2",
|
|
"architectures": [
|
|
"CodeSage"
|
|
],
|
|
"auto_map": {
|
|
"AutoConfig": "config_codesage.CodeSageConfig",
|
|
"AutoTokenizer": "tokenization_codesage.CodeSageTokenizer",
|
|
"AutoModel": "modeling_codesage.CodeSageModel",
|
|
"AutoModelForMaskedLM": "modeling_codesage.CodeSageForMaskedLM",
|
|
"AutoModelForSequenceClassification": "modeling_codesage.CodeSageForSequenceClassification"
|
|
},
|
|
"activation_function": "gelu_new",
|
|
"attention_dropout_prob": 0.1,
|
|
"embedding_dropout_prob": 0.1,
|
|
"initializer_range": 0.02,
|
|
"layer_norm_epsilon": 1e-05,
|
|
"hidden_size": 2048,
|
|
"num_attention_heads": 16,
|
|
"num_hidden_layers": 24,
|
|
"intermediate_size": 8192,
|
|
"max_position_embeddings": 2048,
|
|
"residual_dropout_prob": 0.1,
|
|
"vocab_size": 49154
|
|
} |