{ "architectures": [ "MiniQwen3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "auto_map": { "AutoConfig": "configuration_mini_qwen.MiniQwen3Config", "AutoModelForCausalLM": "modeling_mini_qwen.MiniQwen3ForCausalLM" }, "bos_token_id": 151645, "dtype": "bfloat16", "eos_token_id": 151645, "head_dim": 64, "hidden_act": "silu", "hidden_size": 1024, "initializer_range": 0.02, "intermediate_size": 3072, "max_position_embeddings": 4096, "mlp_bias": false, "model_type": "mini_qwen3", "num_attention_heads": 16, "num_hidden_layers": 24, "num_key_value_heads": 8, "pad_token_id": 151643, "rms_norm_eps": 1e-06, "rope_theta": 1000000.0, "tie_word_embeddings": true, "transformers_version": "5.7.0", "use_cache": true, "use_qk_norm": true, "vocab_size": 151680 }