config.json

{
    "_name_or_path": "/covalent/.cache/models/tmpoel8etc5a80hfd_0",
    "architectures": [
        "MiniCPM3ForCausalLM"
    ],
    "attention_bias": false,
    "attention_dropout": 0.0,
    "auto_map": {
        "AutoConfig": "configuration_minicpm.MiniCPM3Config",
        "AutoModel": "openbmb/MiniCPM3-4B--modeling_minicpm.MiniCPM3Model",
        "AutoModelForCausalLM": "openbmb/MiniCPM3-4B--modeling_minicpm.MiniCPM3ForCausalLM",
        "AutoModelForSeq2SeqLM": "openbmb/MiniCPM3-4B--modeling_minicpm.MiniCPM3ForCausalLM",
        "AutoModelForSequenceClassification": "openbmb/MiniCPM3-4B--modeling_minicpm.MiniCPM3ForSequenceClassification"
    },
    "bos_token_id": 1,
    "dim_model_base": 256,
    "eos_token_id": [
        2,
        73440
    ],
    "head_dim": 96,
    "hidden_act": "silu",
    "hidden_size": 2560,
    "initializer_range": 0.1,
    "intermediate_size": 6400,
    "kv_lora_rank": 256,
    "max_position_embeddings": 32768,
    "model_type": "minicpm3",
    "num_attention_heads": 40,
    "num_hidden_layers": 62,
    "num_key_value_heads": 40,
    "pretraining_tp": 1,
    "q_lora_rank": 768,
    "qk_nope_head_dim": 64,
    "qk_rope_head_dim": 32,
    "quantization_config": {
        "_load_in_4bit": false,
        "_load_in_8bit": true,
        "bnb_4bit_compute_dtype": "bfloat16",
        "bnb_4bit_quant_storage": "uint8",
        "bnb_4bit_quant_type": "fp4",
        "bnb_4bit_use_double_quant": false,
        "llm_int8_enable_fp32_cpu_offload": false,
        "llm_int8_has_fp16_weight": false,
        "llm_int8_skip_modules": [
            "lm_head"
        ],
        "llm_int8_threshold": 6.0,
        "load_in_4bit": false,
        "load_in_8bit": true,
        "quant_method": "bitsandbytes"
    },
    "rms_norm_eps": 1e-05,
    "rope_scaling": {
        "long_factor": [
            1.0591234137867171,
            1.1241891283591912,
            1.2596935748670968,
            1.5380380402321725,
            2.093982484148734,
            3.1446935121267696,
            4.937952647693647,
            7.524541999994549,
            10.475458000005451,
            13.062047352306353,
            14.85530648787323,
            15.906017515851266,
            16.461961959767827,
            16.740306425132907,
            16.87581087164081,
            16.940876586213285
        ],
        "original_max_position_embeddings": 32768,
        "short_factor": [
            1.0591234137867171,
            1.1241891283591912,
            1.2596935748670968,
            1.5380380402321725,
            2.093982484148734,
            3.1446935121267696,
            4.937952647693647,
            7.524541999994549,
            10.475458000005451,
            13.062047352306353,
            14.85530648787323,
            15.906017515851266,
            16.461961959767827,
            16.740306425132907,
            16.87581087164081,
            16.940876586213285
        ],
        "type": "longrope"
    },
    "rope_theta": 10000.0,
    "scale_depth": 1.4,
    "scale_emb": 12,
    "torch_dtype": "float16",
    "transformers_version": "4.46.2",
    "use_cache": true,
    "v_head_dim": 64,
    "vocab_size": 73448,
    "api_key": null
}