36 lines
887 B
JSON
36 lines
887 B
JSON
|
{
|
||
|
|
"architectures": [
|
||
|
|
"MobileLLMForCausalLM"
|
||
|
|
],
|
||
|
|
"attention_bias": false,
|
||
|
|
"attention_dropout": 0.0,
|
||
|
|
"auto_map": {
|
||
|
|
"AutoConfig": "configuration_mobilellm.MobileLLMConfig",
|
||
|
|
"AutoModelForCausalLM": "modeling_mobilellm.MobileLLMForCausalLM"
|
||
|
|
},
|
||
|
|
"bos_token_id": 1,
|
||
|
|
"eos_token_id": 2,
|
||
|
|
"head_dim": 64,
|
||
|
|
"hidden_act": "silu",
|
||
|
|
"hidden_size": 1280,
|
||
|
|
"initializer_range": 0.02,
|
||
|
|
"intermediate_size": 3584,
|
||
|
|
"layer_sharing": false,
|
||
|
|
"max_position_embeddings": 2048,
|
||
|
|
"mlp_bias": false,
|
||
|
|
"model_type": "mobilellm",
|
||
|
|
"num_attention_heads": 20,
|
||
|
|
"num_hidden_layers": 54,
|
||
|
|
"num_key_value_heads": 5,
|
||
|
|
"pretraining_tp": 1,
|
||
|
|
"rms_norm_eps": 1e-05,
|
||
|
|
"rope_scaling": null,
|
||
|
|
"rope_theta": 10000.0,
|
||
|
|
"share_embedding": true,
|
||
|
|
"tie_word_embeddings": false,
|
||
|
|
"torch_dtype": "float16",
|
||
|
|
"transformers_version": "4.41.2",
|
||
|
|
"use_cache": true,
|
||
|
|
"vocab_size": 32000
|
||
|
|
}
|