quantization.md

# Quantization

SGLang supports various quantization methods, including offline quantization and online dynamic quantization.

Offline quantization loads pre-quantized model weights directly during inference. This is required for quantization methods
such as GPTQ and AWQ, which collect and pre-compute various statistics from the original weights using the calibration dataset.

Online quantization dynamically computes scaling parameters—such as the maximum/minimum values of model weights—during runtime.
Like NVIDIA FP8 training's [delayed scaling](https://docs.nvidia.com/deeplearning/transformer-engine/user-guide/examples/fp8_primer.html#Mixed-precision-training-with-FP8) mechanism, online quantization calculates the appropriate scaling factors
on-the-fly to convert high-precision weights into a lower-precision format.

**Note: For better performance, usability and convenience, offline quantization is recommended over online quantization.**

If you use a pre-quantized model, do not add `--quantization` to enable online quantization at the same time.
For popular pre-quantized models, please visit [ModelCloud](https://huggingface.co/collections/ModelCloud/vortex-673743382af0a52b2a8b9fe2)
or [NeuralMagic](https://huggingface.co/collections/neuralmagic) collections on HF for some
popular quality validated quantized models. Quantized models must be validated via benchmarks post-quantization
to guard against abnormal quantization loss regressions.

## Offline Quantization

To load already quantized models, simply load the model weights and config. **Again, if the model has been quantized offline,
there's no need to add `--quantization` argument when starting the engine. The quantization method will be parsed from the
downloaded Hugging Face config. For example, DeepSeek V3/R1 models are already in FP8, so do not add redundant parameters.**

```bash
python3 -m sglang.launch_server \
    --model-path hugging-quants/Meta-Llama-3.1-8B-Instruct-AWQ-INT4 \
    --port 30000 --host 0.0.0.0
```

Take note, if your model is **per-channel quantized (INT8 or FP8) with per-token dynamic quantization activation**, you can opt to include `--quantization w8a8_int8` or `--quantization w8a8_fp8` to invoke the corresponding CUTLASS int8_kernel or fp8_kernel in sgl-kernel. This action will ignore the Hugging Face config's quantization settings. For instance, with `neuralmagic/Meta-Llama-3.1-8B-Instruct-FP8-dynamic`, if you execute with `--quantization w8a8_fp8`, the system will use the `W8A8Fp8Config` from SGLang to invoke the sgl-kernel, rather than the `CompressedTensorsConfig` for vLLM kernels.

```bash
python3 -m sglang.launch_server \
    --model-path neuralmagic/Meta-Llama-3.1-8B-Instruct-FP8-dynamic \
    --quantization w8a8_fp8 \
    --port 30000 --host 0.0.0.0
```

### Examples of Offline Model Quantization


#### Using [auto-round](https://github.com/intel/auto-round)

```bash
# Install
pip install auto-round
```

- LLM quantization

```py
# for LLM
from transformers import AutoModelForCausalLM, AutoTokenizer
from auto_round import AutoRound
model_id = "meta-llama/Llama-3.2-1B-Instruct"
quant_path = "Llama-3.2-1B-Instruct-autoround-4bit"
model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype="auto")
tokenizer = AutoTokenizer.from_pretrained(model_id)
bits, group_size, sym = 4, 128, True # set quantize args
autoround = AutoRound(model, tokenizer, bits=bits, group_size=group_size, sym=sym)
format='auto_round'
autoround.quantize_and_save(quant_path, format=format) # quantize and save

```

- VLM quantization
```py
# for VLMs
from auto_round import AutoRoundMLLM
from transformers import Qwen2VLForConditionalGeneration, AutoProcessor, AutoTokenizer
model_name = "Qwen/Qwen2-VL-2B-Instruct"
quant_path = "Qwen2-VL-2B-Instruct-autoround-4bit"
model = Qwen2VLForConditionalGeneration.from_pretrained(
    model_name, trust_remote_code=True, torch_dtype="auto")
tokenizer = AutoTokenizer.from_pretrained(model_name)
processor = AutoProcessor.from_pretrained(model_name, trust_remote_code=True)
bits, group_size, sym = 4, 128, True
autoround = AutoRoundMLLM(model, tokenizer, processor,
                          bits=bits, group_size=group_size, sym=sym)
format='auto_round'
autoround.quantize_and_save(quant_path, format=format) # quantize and save

```

- Command Line Usage (Gaudi/CPU/Intel GPU/CUDA)

```bash
auto-round \
    --model meta-llama/Llama-3.2-1B-Instruct \
    --bits 4 \
    --group_size 128 \
    --format "auto_gptq,auto_awq,auto_round" \
    --output_dir ./tmp_autoround
```

- known issues

Several limitations currently affect offline quantized model loading in sglang, These issues might be resolved in future updates of sglang. If you experience any problems, consider using Hugging Face Transformers as an alternative.

1. Mixed-bit Quantization Limitations

    Mixed-bit quantization is not fully supported. Due to vLLM's layer fusion (e.g., QKV fusion), applying different bit-widths to components within the same fused layer can lead to compatibility issues.


2. Limited Support for Quantized MoE Models

    Quantized MoE models may encounter inference issues due to kernel limitations (e.g., lack of support for mlp.gate layer quantization). To avoid such errors, please skip quantizing gate layers when processing quantization to MoE modules.


3. Limited Support for Quantized VLMs
    <details>
        <summary>VLM failure cases</summary>

    Qwen2.5-VL-7B

    auto_round:auto_gptq format:  Accuracy is close to zero.

    GPTQ format:  Fails with:
    ```
    The output size is not aligned with the quantized weight shape
    ```

    auto_round:auto_awq and AWQ format:  These work as expected.
    </details>



#### Using [GPTQModel](https://github.com/ModelCloud/GPTQModel)

```bash
# install
pip install gptqmodel --no-build-isolation -v
```

```py
from datasets import load_dataset
from gptqmodel import GPTQModel, QuantizeConfig

model_id = "meta-llama/Llama-3.2-1B-Instruct"
quant_path = "Llama-3.2-1B-Instruct-gptqmodel-4bit"

calibration_dataset = load_dataset(
    "allenai/c4", data_files="en/c4-train.00001-of-01024.json.gz",
    split="train"
  ).select(range(1024))["text"]

quant_config = QuantizeConfig(bits=4, group_size=128) # quantization config
model = GPTQModel.load(model_id, quant_config) # load model

model.quantize(calibration_dataset, batch_size=2) # quantize
model.save(quant_path) # save model
```

#### Using [LLM Compressor](https://github.com/vllm-project/llm-compressor/)

```bash
# install
pip install llmcompressor
```

Here, we take quantize `meta-llama/Meta-Llama-3-8B-Instruct` to `FP8` as an example to elaborate on how to do offline quantization.

```python
from transformers import AutoTokenizer
from llmcompressor.transformers import SparseAutoModelForCausalLM
from llmcompressor.transformers import oneshot
from llmcompressor.modifiers.quantization import QuantizationModifier

# Step 1: Load the original model.
MODEL_ID = "meta-llama/Meta-Llama-3-8B-Instruct"

model = SparseAutoModelForCausalLM.from_pretrained(
  MODEL_ID, device_map="auto", torch_dtype="auto")
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)

# Step 2: Perform offline quantization.
# Step 2.1: Configure the simple PTQ quantization.
recipe = QuantizationModifier(
  targets="Linear", scheme="FP8_DYNAMIC", ignore=["lm_head"])

# Step 2.2: Apply the quantization algorithm.
oneshot(model=model, recipe=recipe)

# Step 3: Save the model.
SAVE_DIR = MODEL_ID.split("/")[1] + "-FP8-Dynamic"
model.save_pretrained(SAVE_DIR)
tokenizer.save_pretrained(SAVE_DIR)
```

Then, you can directly use the quantized model with `SGLang`, by using the following command:

```bash
python3 -m sglang.launch_server \
    --model-path $PWD/Meta-Llama-3-8B-Instruct-FP8-Dynamic \
    --port 30000 --host 0.0.0.0
```

## Online Quantization

To enable online quantization, you can simply specify `--quantization` in the command line. For example, you can launch the server with the following command to enable `FP8` quantization for model `meta-llama/Meta-Llama-3.1-8B-Instruct`:

```bash
python3 -m sglang.launch_server \
    --model-path meta-llama/Meta-Llama-3.1-8B-Instruct \
    --quantization fp8 \
    --port 30000 --host 0.0.0.0
```

Our team is working on supporting more online quantization methods. SGLang will soon support methods including but not limited to `["awq", "gptq", "marlin", "gptq_marlin", "awq_marlin", "bitsandbytes", "gguf"]`.

SGLang also supports quantization methods based on [torchao](https://github.com/pytorch/ao). You can simply specify `--torchao-config` in the command line to support this feature. For example, if you want to enable `int4wo-128` for model `meta-llama/Meta-Llama-3.1-8B-Instruct`, you can launch the server with the following command:

```bash
python3 -m sglang.launch_server \
    --model-path meta-llama/Meta-Llama-3.1-8B-Instruct \
    --torchao-config int4wo-128 \
    --port 30000 --host 0.0.0.0
```

SGLang supports the following quantization methods based on torchao `["int8dq", "int8wo", "fp8wo", "fp8dq-per_tensor", "fp8dq-per_row", "int4wo-32", "int4wo-64", "int4wo-128", "int4wo-256"]`.

Note: According to [this issue](https://github.com/sgl-project/sglang/issues/2219#issuecomment-2561890230), `"int8dq"` method currently has some bugs when using together with cuda graph capture. So we suggest to disable cuda graph capture when using `"int8dq"` method. Namely, please use the following command:

```bash
python3 -m sglang.launch_server \
    --model-path meta-llama/Meta-Llama-3.1-8B-Instruct \
    --torchao-config int8dq \
    --disable-cuda-graph \
    --port 30000 --host 0.0.0.0
```

## Reference

- [GPTQModel](https://github.com/ModelCloud/GPTQModel)
- [LLM Compressor](https://github.com/vllm-project/llm-compressor/)
- [Torchao: PyTorch Architecture Optimization](https://github.com/pytorch/ao)
- [vLLM Quantization](https://docs.vllm.ai/en/latest/quantization/)
- [auto-round](https://github.com/intel/auto-round)