-**Extensive Model Support**: Supports a wide range of generative models (Llama, Gemma, Mistral, QWen, DeepSeek, LLaVA, etc.) and embedding models (e5-mistral), with easy extensibility for integrating new models.
-**Extensive Model Support**: Supports a wide range of generative models (Llama, Gemma, Mistral, QWen, DeepSeek, LLaVA, etc.) and embedding models (e5-mistral), with easy extensibility for integrating new models.
-**Active Community**: SGLang is open-source and backed by an active community with industry adoption.
-**Active Community**: SGLang is open-source and backed by an active community with industry adoption.
Note: Please check the [FlashInfer installation doc](https://docs.flashinfer.ai/installation.html) to install the proper version according to your PyTorch and CUDA versions.
Note: Please check the [FlashInfer installation doc](https://docs.flashinfer.ai/installation.html) to install the proper version according to your PyTorch and CUDA versions.
### Method 3: Using docker
The docker images are available on Docker Hub as [lmsysorg/sglang](https://hub.docker.com/r/lmsysorg/sglang/tags), built from [Dockerfile](https://github.com/sgl-project/sglang/tree/main/docker).
Replace `<secret>` below with your huggingface hub [token](https://huggingface.co/docs/hub/en/security-tokens).
> This method is recommended if you plan to serve it as a service.
> A better approach is to use the [k8s-sglang-service.yaml](docker/k8s-sglang-service.yaml).
1. Copy the [compose.yml](docker/compose.yaml) to your local machine
2. Execute the command `docker compose up -d` in your terminal.
</details>
### Method 5: Run on Kubernetes or Clouds with SkyPilot
<details>
<summary>More</summary>
To deploy on Kubernetes or 12+ clouds, you can use [SkyPilot](https://github.com/skypilot-org/skypilot).
1. Install SkyPilot and set up Kubernetes cluster or cloud access: see [SkyPilot's documentation](https://skypilot.readthedocs.io/en/latest/getting-started/installation.html).
2. Deploy on your own infra with a single command and get the HTTP API endpoint:
3. To further scale up your deployment with autoscaling and failure recovery, check out the [SkyServe + SGLang guide](https://github.com/skypilot-org/skypilot/tree/master/llm/sglang#serving-llama-2-with-sglang-for-more-traffic-using-skyserve).
</details>
### Common Notes
-[FlashInfer](https://github.com/flashinfer-ai/flashinfer) is the default attention kernel backend. It only supports sm75 and above. If you encounter any FlashInfer-related issues on sm75+ devices (e.g., T4, A10, A100, L4, L40S, H100), please switch to other kernels by adding `--attention-backend triton --sampling-backend pytorch` and open an issue on GitHub.
- If you only need to use the OpenAI backend, you can avoid installing other dependencies by using `pip install "sglang[openai]"`.
## Backend: SGLang Runtime (SRT)
## Backend: SGLang Runtime (SRT)
The SGLang Runtime (SRT) is an efficient serving engine.
See [https://sgl-project.github.io/backend.html](https://sgl-project.github.io/backend.html)
{"role":"system","content":"You are a helpful AI assistant"},
{"role":"user","content":"List 3 countries and their capitals."},
],
temperature=0,
max_tokens=64,
)
print(response)
# Text embedding
response=client.embeddings.create(
model="default",
input="How are you today",
)
print(response)
```
It supports streaming, vision, and almost all features of the Chat/Completions/Models/Batch endpoints specified by the [OpenAI API Reference](https://platform.openai.com/docs/api-reference/).
### Additional Server Arguments
- To enable multi-GPU tensor parallelism, add `--tp 2`. If it reports the error "peer access is not supported between these two devices", add `--enable-p2p-check` to the server launch command.
- To enable multi-GPU data parallelism, add `--dp 2`. Data parallelism is better for throughput if there is enough memory. It can also be used together with tensor parallelism. The following command uses 4 GPUs in total.
- If you see out-of-memory errors during serving, try to reduce the memory usage of the KV cache pool by setting a smaller value of `--mem-fraction-static`. The default value is `0.9`.
- To enable the experimental overlapped scheduler, add `--enable-overlap-scheduler`. It overlaps CPU scheduler with GPU computation and can accelerate almost all workloads. This does not work for constrained decoding currenly.
- To enable torch.compile acceleration, add `--enable-torch-compile`. It accelerates small models on small batch sizes. This does not work for FP8 currenly.
- To enable torchao quantization, add `--torchao-config int4wo-128`. It supports various quantization strategies.
- To enable fp8 weight quantization, add `--quantization fp8` on a fp16 checkpoint or directly load a fp8 checkpoint without specifying any arguments.
- To enable fp8 kv cache quantization, add `--kv-cache-dtype fp8_e5m2`.
- If the model does not have a chat template in the Hugging Face tokenizer, you can specify a [custom chat template](docs/custom_chat_template.md).
- To run tensor parallelism on multiple nodes, add `--nnodes 2`. If you have two nodes with two GPUs on each node and want to run TP=4, let `sgl-dev-0` be the hostname of the first node and `50000` be an available port, you can use the following commands. If you meet deadlock, please try to add `--disable-cuda-graph`
- Query the server with the [OpenAI Vision API](https://platform.openai.com/docs/guides/vision). See examples at [test/srt/test_vision_openai_server.py](test/srt/test_vision_openai_server.py)
- Query the server with the [OpenAI Vision API](https://platform.openai.com/docs/guides/vision). See examples at [test/srt/test_vision_openai_server.py](test/srt/test_vision_openai_server.py)
- Benchmark a single static batch by running the following command without launching a server. The arguments are the same as for `launch_server.py`.
Note that this is not a dynamic batching server, so it may run out of memory for a batch size that a real server can handle.
A real server truncates the prefill into several batches, while this unit test does not. For accurate large batch testing, please use `sglang.bench_serving` instead.
## Frontend: Structured Generation Language (SGLang)
## Frontend: Structured Generation Language (SGLang)
The frontend language can be used with local models or API models. It is an alternative to the OpenAI API. You may found it easier to use for complex prompting workflow.
See [https://sgl-project.github.io/frontend.html](https://sgl-project.github.io/frontend.html)
### Quick Start
The example below shows how to use sglang to answer a multi-turn question.
See also [json_decode.py](examples/frontend_language/usage/json_decode.py) for an additional example of specifying formats with Pydantic models.
#### Batching
Use `run_batch` to run a batch of requests with continuous batching.
```python
@sgl.function
deftext_qa(s,question):
s+="Q: "+question+"\n"
s+="A:"+sgl.gen("answer",stop="\n")
states=text_qa.run_batch(
[
{"question":"What is the capital of the United Kingdom?"},
{"question":"What is the capital of France?"},
{"question":"What is the capital of Japan?"},
],
progress_bar=True
)
```
#### Streaming
Add `stream=True` to enable streaming.
```python
@sgl.function
deftext_qa(s,question):
s+="Q: "+question+"\n"
s+="A:"+sgl.gen("answer",stop="\n")
state=text_qa.run(
question="What is the capital of France?",
temperature=0.1,
stream=True
)
foroutinstate.text_iter():
print(out,end="",flush=True)
```
#### Roles
Use `sgl.system`， `sgl.user` and `sgl.assistant` to set roles when using Chat models. You can also define more complex role prompts using begin and end tokens.
```python
@sgl.function
defchat_example(s):
s+=sgl.system("You are a helpful assistant.")
# Same as: s += s.system("You are a helpful assistant.")
withs.user():
s+="Question: What is the capital of France?"
s+=sgl.assistant_begin()
s+="Answer: "+sgl.gen(max_tokens=100,stop="\n")
s+=sgl.assistant_end()
```
#### Tips and Implementation Details
- The `choices` argument in `sgl.gen` is implemented by computing the [token-length normalized log probabilities](https://blog.eleuther.ai/multiple-choice-normalization/) of all choices and selecting the one with the highest probability.
- The `regex` argument in `sgl.gen` is implemented through autoregressive decoding with logit bias masking, according to the constraints set by the regex. It is compatible with `temperature=0` and `temperature != 0`.
## Benchmark And Performance
## Benchmark And Performance
Learn more in our release blogs: [v0.2](https://lmsys.org/blog/2024-07-25-sglang-llama3/), [v0.3](https://lmsys.org/blog/2024-09-04-sglang-v0-3/).
Learn more in our release blogs: [v0.2 blog](https://lmsys.org/blog/2024-07-25-sglang-llama3/), [v0.3 blog](https://lmsys.org/blog/2024-09-04-sglang-v0-3/)
@@ -618,10 +57,3 @@ Learn more in our release blogs: [v0.2](https://lmsys.org/blog/2024-07-25-sglang
...
@@ -618,10 +57,3 @@ Learn more in our release blogs: [v0.2](https://lmsys.org/blog/2024-07-25-sglang
## Citation And Acknowledgment
## Citation And Acknowledgment
Please cite our paper, [SGLang: Efficient Execution of Structured Language Model Programs](https://arxiv.org/abs/2312.07104), if you find the project useful.
Please cite our paper, [SGLang: Efficient Execution of Structured Language Model Programs](https://arxiv.org/abs/2312.07104), if you find the project useful.
We also learned from the design and reused code from the following projects: [Guidance](https://github.com/guidance-ai/guidance), [vLLM](https://github.com/vllm-project/vllm), [LightLLM](https://github.com/ModelTC/lightllm), [FlashInfer](https://github.com/flashinfer-ai/flashinfer), [Outlines](https://github.com/outlines-dev/outlines), and [LMQL](https://github.com/eth-sri/lmql).
We also learned from the design and reused code from the following projects: [Guidance](https://github.com/guidance-ai/guidance), [vLLM](https://github.com/vllm-project/vllm), [LightLLM](https://github.com/ModelTC/lightllm), [FlashInfer](https://github.com/flashinfer-ai/flashinfer), [Outlines](https://github.com/outlines-dev/outlines), and [LMQL](https://github.com/eth-sri/lmql).
- To enable the experimental overlapped scheduler, add `--enable-overlap-scheduler`. It overlaps CPU scheduler with GPU computation and can accelerate almost all workloads. This does not work for constrained decoding currenly.
- To enable torch.compile acceleration, add `--enable-torch-compile`. It accelerates small models on small batch sizes. This does not work for FP8 currenly.
- To enable torch.compile acceleration, add `--enable-torch-compile`. It accelerates small models on small batch sizes. This does not work for FP8 currenly.
- To enable torchao quantization, add `--torchao-config int4wo-128`. It supports various quantization strategies.
- To enable torchao quantization, add `--torchao-config int4wo-128`. It supports various quantization strategies.
- To enable fp8 weight quantization, add `--quantization fp8` on a fp16 checkpoint or directly load a fp8 checkpoint without specifying any arguments.
- To enable fp8 weight quantization, add `--quantization fp8` on a fp16 checkpoint or directly load a fp8 checkpoint without specifying any arguments.
...
@@ -128,7 +129,7 @@ You can view the full example [here](https://github.com/sgl-project/sglang/tree/
...
@@ -128,7 +129,7 @@ You can view the full example [here](https://github.com/sgl-project/sglang/tree/
## Frontend: Structured Generation Language (SGLang)
# Frontend: Structured Generation Language (SGLang)
The frontend language can be used with local models or API models. It is an alternative to the OpenAI API. You may found it easier to use for complex prompting workflow.
The frontend language can be used with local models or API models. It is an alternative to the OpenAI API. You may found it easier to use for complex prompting workflow.
### Quick Start
### Quick Start
The example below shows how to use sglang to answer a mulit-turn question.
The example below shows how to use SGLang to answer a multi-turn question.
