{ "_name_or_path": "bumblebee-testing/tiny-random-Phi3Model", "architectures": [ "Phi3Model" ], "attention_dropout": 0.0, "attention_probs_dropout_prob": 0.1, "bos_token_id": 1, "embd_pdrop": 0.0, "eos_token_id": 32000, "hidden_act": "gelu", "hidden_dropout_prob": 0.1, "hidden_size": 32, "initializer_range": 0.02, "intermediate_size": 37, "max_position_embeddings": 512, "model_type": "phi3", "num_attention_heads": 4, "num_hidden_layers": 2, "num_key_value_heads": 4, "original_max_position_embeddings": 256, "pad_token_id": 0, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": { "long_factor": [ 5.0, 5.0, 5.0, 5.0 ], "short_factor": [ 5.0, 5.0, 5.0, 5.0 ], "type": "longrope" }, "rope_theta": 10000.0, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "float32", "transformers_version": "4.43.4", "type_vocab_size": 16, "use_cache": true, "vocab_size": 1024 }