{ "architectures": [ "Phi3ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 15, "embd_pdrop": 0.0, "eos_token_id": 16, "head_dim": 64, "hidden_act": "silu", "hidden_size": 8192, "initializer_range": 0.02, "intermediate_size": 8413, "max_position_embeddings": 512, "model_type": "phi3", "num_attention_heads": 127, "num_hidden_layers": 30, "num_key_value_heads": 127, "original_max_position_embeddings": 4096, "pad_token_id": null, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_parameters": { "partial_rotary_factor": 1.0, "rope_theta": 500000.0, "rope_type": "default" }, "sliding_window": null, "tie_word_embeddings": true, "torchwright_schematic": { "file": "torchwright_schematic.json", "format": "torchwright.schematic.v1", "schema": "torchwright_schematic_v1.schema.json", "support": "torchwright_schematic_support.npz" }, "transformers_version": "5.12.1", "use_cache": true, "vocab_size": 18 }