README.md

---
library_name: transformers
tags:
- torchao
license: apache-2.0
language:
- multilingual
base_model:
- Qwen/Qwen3-32B
pipeline_tag: text-generation
---

[Qwen3-32B](https://huggingface.co/Qwen3/Qwen3-32B) model quantized with [torchao](https://huggingface.co/docs/transformers/main/en/quantization/torchao) float8 dynamic activation and float8 weight quantization (per row granularity), by PyTorch team. Use it directly, or serve using [vLLM](https://docs.vllm.ai/en/latest/) with TODO VRAM reduction, TODO speedup and little to no accuracy impact on H100.

# 1. Inference with vLLM
```Shell
VLLM_DISABLE_COMPILE_CACHE=1 vllm serve SocialLocalMobile/Qwen3-32B-float8dq --tokenizer Qwen/Qwen3-32B -O3
```

```Shell
curl http://localhost:8000/v1/chat/completions -H "Content-Type: application/json" -d '{
  "model": "SocialLocalMobile/Qwen3-32B-float8dq",
  "messages": [
    {"role": "user", "content": "Give me a short introduction to large language models."}
  ],
  "temperature": 0.6,
  "top_p": 0.95,
  "top_k": 20,
  "max_tokens": 32768
}'
```


# 2. Inference with Transformers
TODO

# 3. Quantization Recipe

Install the required packages:

```Shell
pip install git+https://github.com/huggingface/transformers@main
pip install --pre torchao --index-url https://download.pytorch.org/whl/nightly/cu126
pip install torch
pip install accelerate
```

Use the following code to get the quantized model:

```Py
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, TorchAoConfig

model_id = "Qwen/Qwen3-32B"

## Step 1: Convert to float8
from torchao.quantization import Float8DynamicActivationFloat8WeightConfig, PerRow
quant_config = Float8DynamicActivationFloat8WeightConfig(granularity=PerRow())
quantization_config = TorchAoConfig(quant_type=quant_config)
quantized_model = AutoModelForCausalLM.from_pretrained(
    model_id, 
    device_map="auto", 
    torch_dtype=torch.bfloat16, 
    quantization_config=quantization_config,
)
tokenizer = AutoTokenizer.from_pretrained(model_id)

## Step 2: Sanity check
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("cuda")

# conduct text completion
generated_ids = quantized_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)


# Step 3: Upload to HF
USER_ID = "YOUR_USER_ID"
MODEL_NAME = model_id.split("/")[-1]
save_to = f"{USER_ID}/{MODEL_NAME}-float8dq"
quantized_model.push_to_hub(save_to, safe_serialization=False)
tokenizer.push_to_hub(save_to)
```

# 4. Model Quality
TODO

# 5. Peak Memory Usage
TODO

# 6. Model Performance
TODO

# 7. Disclaimer
PyTorch has not performed safety evaluations or red teamed the quantized models. Performance characteristics, outputs, and behaviors may differ from the original models. Users are solely responsible for selecting appropriate use cases, evaluating and mitigating for accuracy, safety, and fairness, ensuring security, and complying with all applicable laws and regulations.

Nothing contained in this Model Card should be interpreted as or deemed a restriction or modification to the licenses the models are released under, including any limitations of liability or disclaimers of warranties provided therein.