Noumaan/pythonGPT-Sharded
116
1{2 "_name_or_path": "microsoft/Phi-3-mini-4k-instruct",3 "architectures": [4 "Phi3ForCausalLM"5 ],6 "attention_bias": false,7 "attention_dropout": 0.0,8 "auto_map": {9 "AutoConfig": "configuration_phi3.Phi3Config",10 "AutoModel": "modeling_phi3.Phi3ForCausalLM",11 "AutoModelForCausalLM": "microsoft/Phi-3-mini-4k-instruct--modeling_phi3.Phi3ForCausalLM"12 },13 "bos_token_id": 1,14 "embd_pdrop": 0.0,15 "eos_token_id": 32000,16 "hidden_act": "silu",17 "hidden_size": 3072,18 "initializer_range": 0.02,19 "intermediate_size": 8192,20 "max_position_embeddings": 4096,21 "model_type": "phi3",22 "num_attention_heads": 32,23 "num_hidden_layers": 32,24 "num_key_value_heads": 32,25 "original_max_position_embeddings": 4096,26 "pad_token_id": 32000,27 "resid_pdrop": 0.0,28 "rms_norm_eps": 1e-05,29 "rope_scaling": null,30 "rope_theta": 10000.0,31 "sliding_window": 2047,32 "tie_word_embeddings": false,33 "torch_dtype": "bfloat16",34 "transformers_version": "4.41.2",35 "use_cache": true,36 "vocab_size": 3206437}38 