{
    "_name_or_path": "openbmb/CPM-2B",
    "architectures": [
        "MiniCPMForCausalLM"
    ],
    "auto_map": {
        "AutoConfig": "configuration_minicpm.MiniCPMConfig",
        "AutoModel": "modeling_minicpm.MiniCPMModel",
        "AutoModelForCausalLM": "modeling_minicpm.MiniCPMForCausalLM",
        "AutoModelForSeq2SeqLM": "modeling_minicpm.MiniCPMForCausalLM",
        "AutoModelForSequenceClassification": "modeling_minicpm.MiniCPMForSequenceClassification"
    },
    "bos_token_id": 1,
    "eos_token_id": 2,
    "hidden_act": "silu",
    "hidden_size": 1792,
    "initializer_range": 0.1,
    "intermediate_size": 5760,
    "max_position_embeddings": 4096,
    "num_attention_heads": 14,
    "num_hidden_layers": 35,
    "num_key_value_heads": 32,
    "rms_norm_eps": 1e-05,
    "rope_scaling": null,
    "torch_dtype": "bfloat16",
    "transformers_version": "4.36.0",
    "use_cache": true,
    "vocab_size": 122753,
    "scale_emb": 12,
    "dim_model_base": 256,
    "scale_depth": 1.4,
    "tie_word_embeddings": false,
    "norm_after_router": "rms",
    "norm_scale": 1.0,
    "attention_type": "mla",
    "q_lora_rank": 768,
    "kv_lora_rank": 256,
    "qk_nope_head_dim": 128,
    "qk_rope_head_dim": 64,
    "v_head_dim": 128,
    "ffn_type": "block_linear",
    "ffn_gated": true,
    "router_act": "relu",
    "expert_size": 128,
    "num_experts": 56,
    "block_implementation": "torch"
}