Update README.md

README.md

@@ -1,7 +1,6 @@
 ---
 library_name: transformers
 license: apache-2.0
-license_link: https://huggingface.co/Qwen/Qwen3-32B/blob/main/LICENSE
 pipeline_tag: text-generation
 base_model: Qwen/Qwen3-32B
 tags:
@@ -10,94 +9,452 @@ tags:
 - llm-compressor
 ---
 
-# Qwen3-32B W8A8 Quantized
+```
+    ██╗    ██╗ █████╗   █████╗  █████╗ 
+    ██║    ██║██╔══██╗ ██╔══██╗██╔══██╗
+    ██║ █╗ ██║╚█████╔╝ ███████║╚█████╔╝
+    ██║███╗██║██╔══██╗ ██╔══██║██╔══██╗
+    ╚███╔███╔╝╚█████╔╝ ██║  ██║╚█████╔╝
+     ╚══╝╚══╝  ╚════╝  ╚═╝  ╚═╝ ╚════╝ 
+        🗜️ COMPRESSED & OPTIMIZED 🚀
+```
+
+# Qwen3-32B - W8A8 Quantized
+
+W8A8 (8-bit weights and activations) quantized version of Qwen/Qwen3-32B using **LLM-Compressor**.
+
+- 🗜️ **Memory**: ~50% reduction vs FP16  
+- 🚀 **Speed**: Faster inference on supported hardware  
+- 🔗 **Original model**: https://huggingface.co/Qwen/Qwen3-32B 
+- 🏗️ **Recommended architectures**: Ampere and older
+
+<details>
+<summary>Click to view compression config</summary>
+
+```python
+from datasets import load_dataset
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+from llmcompressor.modifiers.quantization import GPTQModifier
+from llmcompressor.modifiers.smoothquant import SmoothQuantModifier
+from llmcompressor.transformers import oneshot
+from llmcompressor.utils import dispatch_for_generation
+
+# Select model and load it.
+model_id = "Qwen/Qwen3-32B"
+model = AutoModelForCausalLM.from_pretrained(
+    model_id,
+    torch_dtype="auto",
+    device_map="auto",
+    low_cpu_mem_usage=True,
+    offload_folder="./offload_tmp",  # Add offload directory
+    max_memory={0: "22GB", 1: "22GB", "cpu": "64GB"},
+)
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+
+# Select calibration dataset.
+DATASET_ID = "HuggingFaceH4/ultrachat_200k"
+DATASET_SPLIT = "train_sft"
+
+# Select number of samples. 512 samples is a good place to start.
+# Increasing the number of samples can improve accuracy.
+NUM_CALIBRATION_SAMPLES = 512
+MAX_SEQUENCE_LENGTH = 2048
+
+# Load dataset and preprocess.
+ds = load_dataset(DATASET_ID, split=f"{DATASET_SPLIT}[:{NUM_CALIBRATION_SAMPLES}]")
+ds = ds.shuffle(seed=42)
+
+
+def preprocess(example):
+    return {
+        "text": tokenizer.apply_chat_template(
+            example["messages"],
+            tokenize=False,
+        )
+    }
+
+
+ds = ds.map(preprocess)
+
+
+# Tokenize inputs.
+def tokenize(sample):
+    return tokenizer(
+        sample["text"],
+        padding=False,
+        max_length=MAX_SEQUENCE_LENGTH,
+        truncation=True,
+        add_special_tokens=False,
+    )
+
+
+ds = ds.map(tokenize, remove_columns=ds.column_names)
+
+# Configure the quantization algorithm to run.
+#   * apply SmoothQuant to make the activations easier to quantize
+#   * quantize the weights to int8 with GPTQ (static per channel)
+#   * quantize the activations to int8 (dynamic per token)
+recipe = [
+    SmoothQuantModifier(smoothing_strength=0.8),
+    GPTQModifier(targets="Linear", scheme="W8A8", ignore=["lm_head"]),
+]
+# Apply algorithms.
+oneshot(
+    model=model,
+    dataset=ds,
+    recipe=recipe,
+    max_seq_length=MAX_SEQUENCE_LENGTH,
+    num_calibration_samples=NUM_CALIBRATION_SAMPLES,
+    output_dir="./Qwen3-32B-W8A8",  # Add this line
+)
+
+# Save to disk compressed.
+SAVE_DIR = model_id.rstrip("/").split("/")[-1] + "-W8A8"
+model.save_pretrained(SAVE_DIR, save_compressed=True)
+tokenizer.save_pretrained(SAVE_DIR)
+```
+
+</details>
 
-This is a W8A8 quantized version of [Qwen/Qwen3-32B](https://huggingface.co/Qwen/Qwen3-32B) using [LLM-Compressor](https://github.com/vllm-project/llm-compressor).
+---
+
+## 📄 Original Model README
+
+# Qwen3-32B
+<a href="https://chat.qwen.ai/" target="_blank" style="margin: 2px;">
+    <img alt="Chat" src="https://img.shields.io/badge/%F0%9F%92%9C%EF%B8%8F%20Qwen%20Chat%20-536af5" style="display: inline-block; vertical-align: middle;"/>
+</a>
+
+## Qwen3 Highlights
 
-## Quantization Details
+Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models. Built upon extensive training, Qwen3 delivers groundbreaking advancements in reasoning, instruction-following, agent capabilities, and multilingual support, with the following key features:
 
-- **Base Model**: Qwen/Qwen3-32B
-- **Quantization Method**: W8A8 (8-bit weights, 8-bit activations)
-- **Quantization Framework**: LLM-Compressor
-- **Model Size**: Significantly reduced from original 32.8B parameters
-- **Precision**: INT8 for both weights and activations
+- **Uniquely support of seamless switching between thinking mode** (for complex logical reasoning, math, and coding) and **non-thinking mode** (for efficient, general-purpose dialogue) **within single model**, ensuring optimal performance across various scenarios.
+- **Significantly enhancement in its reasoning capabilities**, surpassing previous QwQ (in thinking mode) and Qwen2.5 instruct models (in non-thinking mode) on mathematics, code generation, and commonsense logical reasoning.
+- **Superior human preference alignment**, excelling in creative writing, role-playing, multi-turn dialogues, and instruction following, to deliver a more natural, engaging, and immersive conversational experience.
+- **Expertise in agent capabilities**, enabling precise integration with external tools in both thinking and unthinking modes and achieving leading performance among open-source models in complex agent-based tasks.
+- **Support of 100+ languages and dialects** with strong capabilities for **multilingual instruction following** and **translation**.
 
-## Performance Considerations
+## Model Overview
 
-- **Memory Usage**: Significantly reduced memory footprint compared to the original FP16/BF16 model
-- **Inference Speed**: Faster inference due to reduced precision and smaller model size
-- **Accuracy**: Minimal accuracy loss compared to the original model (specific benchmarks may vary)
+**Qwen3-32B** has the following features:
+- Type: Causal Language Models
+- Training Stage: Pretraining & Post-training
+- Number of Parameters: 32.8B
+- Number of Paramaters (Non-Embedding): 31.2B
+- Number of Layers: 64
+- Number of Attention Heads (GQA): 64 for Q and 8 for KV
+- Context Length: 32,768 natively and [131,072 tokens with YaRN](#processing-long-texts). 
 
-## Hardware Requirements
+For more details, including benchmark evaluation, hardware requirements, and inference performance, please refer to our [blog](https://qwenlm.github.io/blog/qwen3/), [GitHub](https://github.com/QwenLM/Qwen3), and [Documentation](https://qwen.readthedocs.io/en/latest/).
 
-This quantized model has lower hardware requirements than the original:
-- **Memory**: Approximately 2x less GPU memory required
-- **Compute**: Compatible with INT8 tensor operations
-- **Recommended**: GPUs with tensor core support for optimal INT8 performance
+## Quickstart
+
+The code of Qwen3 has been in the latest Hugging Face `transformers` and we advise you to use the latest version of `transformers`.
+
+With `transformers<4.51.0`, you will encounter the following error:
+```
+KeyError: 'qwen3'
+```
 
-## All Original Features Preserved
+The following contains a code snippet illustrating how to use the model generate content based on given inputs. 
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
 
-This quantized model retains all the capabilities of the original Qwen3-32B:
+model_name = "Qwen/Qwen3-32B"
 
-- **Thinking Mode Support**: Seamless switching between thinking and non-thinking modes
-- **Enhanced Reasoning**: Superior performance in mathematics, code generation, and logical reasoning
-- **Multilingual Support**: 100+ languages and dialects
-- **Agent Capabilities**: Tool calling and external integration
-- **Long Context**: Native 32,768 token support, extensible to 131,072 with YaRN
+# load the tokenizer and the model
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    torch_dtype="auto",
+    device_map="auto"
+)
+
+# prepare the model input
+prompt = "Give me a short introduction to large language model."
+messages = [
+    {"role": "user", "content": prompt}
+]
+text = tokenizer.apply_chat_template(
+    messages,
+    tokenize=False,
+    add_generation_prompt=True,
+    enable_thinking=True # Switches between thinking and non-thinking modes. Default is True.
+)
+model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
+
+# conduct text completion
+generated_ids = model.generate(
+    **model_inputs,
+    max_new_tokens=32768
+)
+output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist() 
+
+# parsing thinking content
+try:
+    # rindex finding 151668 (</think>)
+    index = len(output_ids) - output_ids[::-1].index(151668)
+except ValueError:
+    index = 0
+
+thinking_content = tokenizer.decode(output_ids[:index], skip_special_tokens=True).strip("\n")
+content = tokenizer.decode(output_ids[index:], skip_special_tokens=True).strip("\n")
+
+print("thinking content:", thinking_content)
+print("content:", content)
+```
+
+For deployment, you can use `sglang>=0.4.6.post1` or `vllm>=0.8.5` or to create an OpenAI-compatible API endpoint:
+- SGLang:
+    ```shell
+    python -m sglang.launch_server --model-path Qwen/Qwen3-32B --reasoning-parser qwen3
+    ```
+- vLLM:
+    ```shell
+    vllm serve Qwen/Qwen3-32B --enable-reasoning --reasoning-parser deepseek_r1
+    ```
+
+For local use, applications such as Ollama, LMStudio, MLX-LM, llama.cpp, and KTransformers have also supported Qwen3.
 
 ## Switching Between Thinking and Non-Thinking Mode
 
-The quantized model supports the same thinking mode controls as the original:
+> [!TIP]
+> The `enable_thinking` switch is also available in APIs created by SGLang and vLLM. 
+> Please refer to our documentation for [SGLang](https://qwen.readthedocs.io/en/latest/deployment/sglang.html#thinking-non-thinking-modes) and [vLLM](https://qwen.readthedocs.io/en/latest/deployment/vllm.html#thinking-non-thinking-modes) users.
+
+### `enable_thinking=True`
+
+By default, Qwen3 has thinking capabilities enabled, similar to QwQ-32B. This means the model will use its reasoning abilities to enhance the quality of generated responses. For example, when explicitly setting `enable_thinking=True` or leaving it as the default value in `tokenizer.apply_chat_template`, the model will engage its thinking mode.
 
-### `enable_thinking=True` (Default)
 ```python
 text = tokenizer.apply_chat_template(
     messages,
     tokenize=False,
     add_generation_prompt=True,
-    enable_thinking=True
+    enable_thinking=True  # True is the default value for enable_thinking
 )
 ```
 
+In this mode, the model will generate think content wrapped in a `<think>...</think>` block, followed by the final response.
+
+> [!NOTE]
+> For thinking mode, use `Temperature=0.6`, `TopP=0.95`, `TopK=20`, and `MinP=0` (the default setting in `generation_config.json`). **DO NOT use greedy decoding**, as it can lead to performance degradation and endless repetitions. For more detailed guidance, please refer to the [Best Practices](#best-practices) section.
+
+
 ### `enable_thinking=False`
+
+We provide a hard switch to strictly disable the model's thinking behavior, aligning its functionality with the previous Qwen2.5-Instruct models. This mode is particularly useful in scenarios where disabling thinking is essential for enhancing efficiency.
+
 ```python
 text = tokenizer.apply_chat_template(
     messages,
     tokenize=False,
     add_generation_prompt=True,
-    enable_thinking=False
+    enable_thinking=False  # Setting enable_thinking=False disables thinking mode
 )
 ```
 
-## Best Practices
+In this mode, the model will not generate any think content and will not include a `<think>...</think>` block.
 
-Follow the same best practices as the original model:
+> [!NOTE]
+> For non-thinking mode, we suggest using `Temperature=0.7`, `TopP=0.8`, `TopK=20`, and `MinP=0`. For more detailed guidance, please refer to the [Best Practices](#best-practices) section.
 
-1. **Sampling Parameters**:
-   - Thinking mode: `Temperature=0.6`, `TopP=0.95`, `TopK=20`, `MinP=0`
-   - Non-thinking mode: `Temperature=0.7`, `TopP=0.8`, `TopK=20`, `MinP=0`
+### Advanced Usage: Switching Between Thinking and Non-Thinking Modes via User Input
+
+We provide a soft switch mechanism that allows users to dynamically control the model's behavior when `enable_thinking=True`. Specifically, you can add `/think` and `/no_think` to user prompts or system messages to switch the model's thinking mode from turn to turn. The model will follow the most recent instruction in multi-turn conversations.
+
+Here is an example of a multi-turn conversation:
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+class QwenChatbot:
+    def __init__(self, model_name="Qwen/Qwen3-32B"):
+        self.tokenizer = AutoTokenizer.from_pretrained(model_name)
+        self.model = AutoModelForCausalLM.from_pretrained(model_name)
+        self.history = []
+
+    def generate_response(self, user_input):
+        messages = self.history + [{"role": "user", "content": user_input}]
+
+        text = self.tokenizer.apply_chat_template(
+            messages,
+            tokenize=False,
+            add_generation_prompt=True
+        )
+
+        inputs = self.tokenizer(text, return_tensors="pt")
+        response_ids = self.model.generate(**inputs, max_new_tokens=32768)[0][len(inputs.input_ids[0]):].tolist()
+        response = self.tokenizer.decode(response_ids, skip_special_tokens=True)
+
+        # Update history
+        self.history.append({"role": "user", "content": user_input})
+        self.history.append({"role": "assistant", "content": response})
+
+        return response
+
+# Example Usage
+if __name__ == "__main__":
+    chatbot = QwenChatbot()
+
+    # First input (without /think or /no_think tags, thinking mode is enabled by default)
+    user_input_1 = "How many r's in strawberries?"
+    print(f"User: {user_input_1}")
+    response_1 = chatbot.generate_response(user_input_1)
+    print(f"Bot: {response_1}")
+    print("----------------------")
+
+    # Second input with /no_think
+    user_input_2 = "Then, how many r's in blueberries? /no_think"
+    print(f"User: {user_input_2}")
+    response_2 = chatbot.generate_response(user_input_2)
+    print(f"Bot: {response_2}") 
+    print("----------------------")
+
+    # Third input with /think
+    user_input_3 = "Really? /think"
+    print(f"User: {user_input_3}")
+    response_3 = chatbot.generate_response(user_input_3)
+    print(f"Bot: {response_3}")
+```
+
+> [!NOTE]
+> For API compatibility, when `enable_thinking=True`, regardless of whether the user uses `/think` or `/no_think`, the model will always output a block wrapped in `<think>...</think>`. However, the content inside this block may be empty if thinking is disabled.
+> When `enable_thinking=False`, the soft switches are not valid. Regardless of any `/think` or `/no_think` tags input by the user, the model will not generate think content and will not include a `<think>...</think>` block.
+
+## Agentic Use
+
+Qwen3 excels in tool calling capabilities. We recommend using [Qwen-Agent](https://github.com/QwenLM/Qwen-Agent) to make the best use of agentic ability of Qwen3. Qwen-Agent encapsulates tool-calling templates and tool-calling parsers internally, greatly reducing coding complexity.
+
+To define the available tools, you can use the MCP configuration file, use the integrated tool of Qwen-Agent, or integrate other tools by yourself.
+```python
+from qwen_agent.agents import Assistant
+
+# Define LLM
+llm_cfg = {
+    'model': 'Qwen3-32B',
 
-2. **Output Length**: Use 32,768 tokens for most queries, 38,912 for complex problems
+    # Use the endpoint provided by Alibaba Model Studio:
+    # 'model_type': 'qwen_dashscope',
+    # 'api_key': os.getenv('DASHSCOPE_API_KEY'),
 
-3. **Avoid Greedy Decoding**: Do not use greedy decoding in thinking mode
+    # Use a custom endpoint compatible with OpenAI API:
+    'model_server': 'http://localhost:8000/v1',  # api_base
+    'api_key': 'EMPTY',
 
+    # Other parameters:
+    # 'generate_cfg': {
+    #         # Add: When the response content is `<think>this is the thought</think>this is the answer;
+    #         # Do not add: When the response has been separated by reasoning_content and content.
+    #         'thought_in_content': True,
+    #     },
+}
+
+# Define Tools
+tools = [
+    {'mcpServers': {  # You can specify the MCP configuration file
+            'time': {
+                'command': 'uvx',
+                'args': ['mcp-server-time', '--local-timezone=Asia/Shanghai']
+            },
+            "fetch": {
+                "command": "uvx",
+                "args": ["mcp-server-fetch"]
+            }
+        }
+    },
+  'code_interpreter',  # Built-in tools
+]
+
+# Define Agent
+bot = Assistant(llm=llm_cfg, function_list=tools)
+
+# Streaming generation
+messages = [{'role': 'user', 'content': 'https://qwenlm.github.io/blog/ Introduce the latest developments of Qwen'}]
+for responses in bot.run(messages=messages):
+    pass
+print(responses)
+```
+
+## Processing Long Texts
+
+Qwen3 natively supports context lengths of up to 32,768 tokens. For conversations where the total length (including both input and output) significantly exceeds this limit, we recommend using RoPE scaling techniques to handle long texts effectively. We have validated the model's performance on context lengths of up to 131,072 tokens using the [YaRN](https://arxiv.org/abs/2309.00071) method.
 
-## Original Model Information
+YaRN is currently supported by several inference frameworks, e.g., `transformers` and `llama.cpp` for local use, `vllm` and `sglang` for deployment. In general, there are two approaches to enabling YaRN for supported frameworks:
 
-For complete documentation, benchmarks, and detailed usage instructions, please refer to the original [Qwen3-32B model card](https://huggingface.co/Qwen/Qwen3-32B).
+- Modifying the model files:
+  In the `config.json` file, add the `rope_scaling` fields:
+    ```json
+    {
+        ...,
+        "rope_scaling": {
+            "rope_type": "yarn",
+            "factor": 4.0,
+            "original_max_position_embeddings": 32768
+        }
+    }
+    ```
+  For `llama.cpp`, you need to regenerate the GGUF file after the modification.
 
-### Key Specifications (from original model):
-- **Type**: Causal Language Models
-- **Parameters**: 32.8B total, 31.2B non-embedding
-- **Layers**: 64
-- **Attention Heads**: 64 for Q, 8 for KV (GQA)
-- **Context Length**: 32,768 tokens natively, 131,072 with YaRN
+- Passing command line arguments:
 
-## Citation
+  For `vllm`, you can use
+    ```shell
+    vllm serve ... --rope-scaling '{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":32768}' --max-model-len 131072  
+    ```
 
-If you use this quantized model, please cite the original Qwen3 work.
+  For `sglang`, you can use
+    ```shell
+    python -m sglang.launch_server ... --json-model-override-args '{"rope_scaling":{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":32768}}'
+    ```
+
+  For `llama-server` from `llama.cpp`, you can use
+    ```shell
+    llama-server ... --rope-scaling yarn --rope-scale 4 --yarn-orig-ctx 32768
+    ```
+
+> [!IMPORTANT]
+> If you encounter the following warning
+> ```
+> Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'original_max_position_embeddings'}
+> ```
+> please upgrade `transformers>=4.51.0`.
+
+> [!NOTE]
+> All the notable open-source frameworks implement static YaRN, which means the scaling factor remains constant regardless of input length, **potentially impacting performance on shorter texts.**
+> We advise adding the `rope_scaling` configuration only when processing long contexts is required. 
+> It is also recommended to modify the `factor` as needed. For example, if the typical context length for your application is 65,536 tokens, it would be better to set `factor` as 2.0. 
+
+> [!NOTE]
+> The default `max_position_embeddings` in `config.json` is set to 40,960. This allocation includes reserving 32,768 tokens for outputs and 8,192 tokens for typical prompts, which is sufficient for most scenarios involving short text processing. If the average context length does not exceed 32,768 tokens, we do not recommend enabling YaRN in this scenario, as it may potentially degrade model performance.
+
+> [!TIP]
+> The endpoint provided by Alibaba Model Studio supports dynamic YaRN by default and no extra configuration is needed.
+
+## Best Practices
 
-```bibtex
+To achieve optimal performance, we recommend the following settings:
+
+1. **Sampling Parameters**:
+   - For thinking mode (`enable_thinking=True`), use `Temperature=0.6`, `TopP=0.95`, `TopK=20`, and `MinP=0`. **DO NOT use greedy decoding**, as it can lead to performance degradation and endless repetitions.
+   - For non-thinking mode (`enable_thinking=False`), we suggest using `Temperature=0.7`, `TopP=0.8`, `TopK=20`, and `MinP=0`.
+   - For supported frameworks, you can adjust the `presence_penalty` parameter between 0 and 2 to reduce endless repetitions. However, using a higher value may occasionally result in language mixing and a slight decrease in model performance.
+
+2. **Adequate Output Length**: We recommend using an output length of 32,768 tokens for most queries. For benchmarking on highly complex problems, such as those found in math and programming competitions, we suggest setting the max output length to 38,912 tokens. This provides the model with sufficient space to generate detailed and comprehensive responses, thereby enhancing its overall performance.
+
+3. **Standardize Output Format**: We recommend using prompts to standardize model outputs when benchmarking.
+   - **Math Problems**: Include "Please reason step by step, and put your final answer within \boxed{}." in the prompt.
+   - **Multiple-Choice Questions**: Add the following JSON structure to the prompt to standardize responses: "Please show your choice in the `answer` field with only the choice letter, e.g., `"answer": "C"`."
+
+4. **No Thinking Content in History**: In multi-turn conversations, the historical model output should only include the final output part and does not need to include the thinking content. It is implemented in the provided chat template in Jinja2. However, for frameworks that do not directly use the Jinja2 chat template, it is up to the developers to ensure that the best practice is followed.
+
+### Citation
+
+If you find our work helpful, feel free to give us a cite.
+
+```
 @misc{qwen3technicalreport,
       title={Qwen3 Technical Report}, 
       author={Qwen Team},
@@ -107,9 +464,4 @@ If you use this quantized model, please cite the original Qwen3 work.
       primaryClass={cs.CL},
       url={https://arxiv.org/abs/2505.09388}, 
 }
-
-```
-
-## Disclaimer
-
-This is an unofficial quantized version. For the official model and support, please refer to the original [Qwen3-32B](https://huggingface.co/Qwen/Qwen3-32B) repository.
+```