38 lines
1.1 KiB
JSON
38 lines
1.1 KiB
JSON
{
|
|
"_name_or_path": "mini_cpm_russ",
|
|
"architectures": [
|
|
"MiniCPMForCausalLM"
|
|
],
|
|
"attention_bias": false,
|
|
"attention_dropout": 0.0,
|
|
"auto_map": {
|
|
"AutoConfig": "configuration_minicpm.MiniCPMConfig",
|
|
"AutoModel": "openbmb/MiniCPM-2B-sft-bf16--modeling_minicpm.MiniCPMModel",
|
|
"AutoModelForCausalLM": "modeling_minicpm.MiniCPMForCausalLM",
|
|
"AutoModelForSeq2SeqLM": "openbmb/MiniCPM-2B-sft-bf16--modeling_minicpm.MiniCPMForCausalLM",
|
|
"AutoModelForSequenceClassification": "openbmb/MiniCPM-2B-sft-bf16--modeling_minicpm.MiniCPMForSequenceClassification"
|
|
},
|
|
"bos_token_id": 1,
|
|
"dim_model_base": 256,
|
|
"eos_token_id": 2,
|
|
"hidden_act": "silu",
|
|
"hidden_size": 2304,
|
|
"initializer_range": 0.1,
|
|
"intermediate_size": 5760,
|
|
"max_position_embeddings": 2048,
|
|
"model_type": "minicpm",
|
|
"num_attention_heads": 36,
|
|
"num_hidden_layers": 40,
|
|
"num_key_value_heads": 36,
|
|
"pretraining_tp": 1,
|
|
"rms_norm_eps": 1e-05,
|
|
"rope_scaling": null,
|
|
"rope_theta": 10000.0,
|
|
"scale_depth": 1.4,
|
|
"scale_emb": 12,
|
|
"torch_dtype": "float16",
|
|
"transformers_version": "4.39.0.dev0",
|
|
"use_cache": true,
|
|
"vocab_size": 148106
|
|
}
|