{ "architectures": [ "Phi3ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 0, "dtype": "float32", "embd_pdrop": 0.0, "eos_token_id": 0, "hidden_act": "silu", "hidden_size": 768, "ignore_keys_at_rope_validation": null, "initializer_range": 0.02, "intermediate_size": 2304, "max_position_embeddings": 1024, "model_type": "phi3", "num_attention_heads": 6, "num_hidden_layers": 24, "num_key_value_heads": 6, "original_max_position_embeddings": 4096, "pad_token_id": 4, "partial_rotary_factor": 1.0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_parameters": { "partial_rotary_factor": 1.0, "rope_theta": 10000.0, "rope_type": "default" }, "sliding_window": null, "tie_word_embeddings": false, "transformers_version": "5.2.0", "use_cache": false, "vocab_size": 32000 }