config.json

{
  "architectures": [
    "Qwen3ForCausalLM"
  ],
  "auto_map": {
    "AutoModelForCausalLM": "Qwen3ForCausalLM"
  },
  "model_type": "qwen3",
  "quantization_config": {
    "quant_method": "gguf",
    "bits": 8,
    "group_size": 32,
    "zero_point": true,
    "desc_act": false
  },
  "gguf_metadata": {
    "general": {
      "name": "Qwen3 4B Function Calling Pro",
      "architecture": "qwen3",
      "file_type": 7,
      "quantization_version": 2,
      "size_label": "4B"
    },
    "qwen3": {
      "context_length": 262144,
      "embedding_length": 2560,
      "block_count": 36,
      "feed_forward_length": 9728,
      "attention": {
        "head_count": 32,
        "head_count_kv": 8,
        "key_length": 128,
        "value_length": 128,
        "layer_norm_rms_epsilon": 1e-06
      },
      "rope": {
        "freq_base": 5000000.0
      }
    }
  },
  "vocab_size": 151936,
  "hidden_size": 2560,
  "intermediate_size": 9728,
  "num_hidden_layers": 36,
  "num_attention_heads": 32,
  "num_key_value_heads": 8,
  "hidden_act": "silu",
  "max_position_embeddings": 262144,
  "initializer_range": 0.02,
  "rms_norm_eps": 1e-06,
  "use_cache": true,
  "rope_theta": 5000000.0,
  "attention_bias": false,
  "attention_dropout": 0.0,
  "torch_dtype": "float16",
  "transformers_version": "4.45.0",
  "special_tokens": {
    "bos_token": "<|endoftext|>",
    "eos_token": "<|im_end|>",
    "pad_token": "<|im_end|>",
    "additional_special_tokens": [
      "<|im_start|>",
      "<|im_end|>",
      "<|object_ref_start|>",
      "<|object_ref_end|>",
      "<|box_start|>",
      "<|box_end|>",
      "<|quad_start|>",
      "<|quad_end|>",
      "<|vision_start|>",
      "<|vision_end|>",
      "<|vision_pad|>",
      "<|image_pad|>",
      "<|video_pad|>",
      "<tool_call>",
      "</tool_call>",
      "<tool_response>",
      "</tool_response>"
    ]
  },
  "chat_template": "{% for message in messages %}{% if message['role'] == 'user' %}{{ '<start_of_turn>user\n' + message['content'] + '<end_of_turn>\n' }}{% elif message['role'] == 'system' %}{{ '<start_of_turn>system\n' + message['content'] + '<end_of_turn>\n' }}{% elif message['role'] == 'assistant' %}{{ '<start_of_turn>model\n' + message['content'] + '<end_of_turn>\n' }}{% endif %}{% if loop.last and add_generation_prompt %}{{ '<start_of_turn>model\n' }}{% endif %}{% endfor %}",
  "generation_config": {
    "do_sample": true,
    "temperature": 0.7,
    "top_p": 0.8,
    "top_k": 40,
    "repetition_penalty": 1.1,
    "max_new_tokens": 512,
    "stop": [
      "<|im_end|>",
      "<|im_start|>",
      "<tool_call>",
      "</tool_call>",
      "<tool_response>",
      "</tool_response>"
    ]
  }
}