openbmb/MiniCPM-2B-128k/bert4torch_config.json

{
    "bos_token_id": 1,
    "eos_token_id": 2,
    "hidden_act": "silu",
    "hidden_size": 2304,
    "initializer_range": 0.1,
    "intermediate_size": 5760,
    "max_position_embeddings": 65536,
    "num_attention_heads": 36,
    "num_hidden_layers": 40,
    "num_key_value_heads": 36,
    "layer_norm_eps": 1e-05,
    "rope_scaling": {"type": "dynamic", "factor": 4.0},
    "torch_dtype": "bfloat16",
    "vocab_size": 122760,
    "emb_scale": 12,
    "dim_model_base": 256,
    "scale_depth": 1.4,
    "tie_word_embeddings": false,
	"rope_rank": "updown",
    "rope_theta": 1000000.0,
	"segment_vocab_size": 0,
	"skip_init": true,
	"model": "minicpm",
    "template": "llama3",
    "convert_lm_logits_dtype": "float32",
	"generation_config": {"tokenizer_config":  {"skip_special_tokens": true, "add_special_tokens": true}, 
                          "eos_token_id": [2], "max_length": 131072}
}