{"architectures": ["MistralForCausalLM"], "attention_dropout": 0.0, "bos_token_id": 0, "eos_token_id": 0, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "intermediate_size": 5632, "max_position_embeddings": 4096, "model_type": "mistral", "neuron": {"auto_cast_type": "fp16", "batch_size": 4, "checkpoint_id": "optimum/mistral-en-1.1B-testing", "checkpoint_revision": "df7d850906b071dbb69cf11af1033d7cd735227d", "compiler_type": "neuronx-cc", "compiler_version": "2.13.66.0+6dfecc895", "num_cores": 2, "sequence_length": 2048, "task": "text-generation"}, "num_attention_heads": 32, "num_hidden_layers": 22, "num_key_value_heads": 4, "rms_norm_eps": 1e-05, "rope_theta": 10000.0, "sliding_window": 1024, "tie_word_embeddings": false, "torch_dtype": "float16", "use_cache": true, "vocab_size": 50257}