33 lines
854 B
JSON
33 lines
854 B
JSON
{
|
|
"_name_or_path": "Index_1_9B",
|
|
"architectures": [
|
|
"IndexForCausalLM"
|
|
],
|
|
"auto_map": {
|
|
"AutoConfig": "configuration_index.IndexConfig",
|
|
"AutoModelForCausalLM": "modeling_index.IndexForCausalLM",
|
|
"AutoModelForSequenceClassification": "modeling_index.IndexForSequenceClassification"
|
|
},
|
|
"bos_token_id": 1,
|
|
"eos_token_id": 2,
|
|
"pad_token_id": 0,
|
|
"hidden_act": "silu",
|
|
"hidden_size": 2048,
|
|
"initializer_range": 0.01,
|
|
"intermediate_size": 5888,
|
|
"max_length": 4096,
|
|
"max_position_embeddings": 4096,
|
|
"model_type": "index",
|
|
"num_attention_heads": 16,
|
|
"num_key_value_heads": 16,
|
|
"num_hidden_layers": 36,
|
|
"rms_norm_eps": 1e-06,
|
|
"rope_scaling": null,
|
|
"tie_word_embeddings": false,
|
|
"norm_head":1,
|
|
"torch_dtype": "bfloat16",
|
|
"transformers_version": "4.39.2",
|
|
"use_cache": true,
|
|
"vocab_size": 65029
|
|
}
|