Edcastro/MiniCPM-1B-sft-bf16-edcastr_JavaScript-v2
030
1{2 "architectures": [3 "MiniCPMForCausalLM"4 ],5 "attention_bias": false,6 "attention_dropout": 0.0,7 "auto_map": {8 "AutoConfig": "configuration_minicpm.MiniCPMConfig",9 "AutoModel": "modeling_minicpm.MiniCPMModel",10 "AutoModelForCausalLM": "modeling_minicpm.MiniCPMForCausalLM",11 "AutoModelForSeq2SeqLM": "modeling_minicpm.MiniCPMForCausalLM",12 "AutoModelForSequenceClassification": "modeling_minicpm.MiniCPMForSequenceClassification"13 },14 "bos_token_id": 1,15 "dim_model_base": 256,16 "dtype": "bfloat16",17 "eos_token_id": 2,18 "hidden_act": "silu",19 "hidden_size": 1536,20 "initializer_range": 0.1,21 "intermediate_size": 3840,22 "max_position_embeddings": 4096,23 "model_type": "minicpm",24 "num_attention_heads": 24,25 "num_hidden_layers": 52,26 "num_key_value_heads": 8,27 "pretraining_tp": 1,28 "rms_norm_eps": 1e-05,29 "rope_scaling": null,30 "rope_theta": 10000.0,31 "scale_depth": 1.4,32 "scale_emb": 12,33 "transformers_version": "4.57.0",34 "use_cache": true,35 "vocab_size": 7344036}37 