{ "architectures": [ "ILLaDAForCausalLM" ], "auto_map": { "AutoConfig": "configuration_illada.ILLaDAConfig", "AutoModel": "modeling_illada.ILLaDAForCausalLM", "AutoModelForCausalLM": "modeling_illada.ILLaDAForCausalLM" }, "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 0, "dtype": "bfloat16", "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.013975424859373685, "intermediate_size": 14336, "layer_norm_eps": null, "max_position_embeddings": 8192, "mlp_bias": false, "model_type": "illada", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 8, "resid_pdrop": 0.0, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 1.0, "rope_type": "default" }, "rope_theta": 10000.0, "tie_word_embeddings": true, "transformers_version": "4.57.1", "torch_dtype": "bfloat16", "vocab_size": 155136 }