README.md 22.7 KB
Newer Older
Lianmin Zheng's avatar
Lianmin Zheng committed
1
<div align="center">
2
<img src="https://raw.githubusercontent.com/sgl-project/sglang/main/assets/logo.png" alt="logo" width="400"></img>
Lianmin Zheng's avatar
Lianmin Zheng committed
3

Yineng Zhang's avatar
Yineng Zhang committed
4
5
6
7
8
9
[![PyPI](https://img.shields.io/pypi/v/sglang)](https://pypi.org/project/sglang)
![PyPI - Downloads](https://img.shields.io/pypi/dm/sglang)
[![license](https://img.shields.io/github/license/sgl-project/sglang.svg)](https://github.com/sgl-project/sglang/tree/main/LICENSE)
[![issue resolution](https://img.shields.io/github/issues-closed-raw/sgl-project/sglang)](https://github.com/sgl-project/sglang/issues)
[![open issues](https://img.shields.io/github/issues-raw/sgl-project/sglang)](https://github.com/sgl-project/sglang/issues)

Yineng Zhang's avatar
Yineng Zhang committed
10
11
</div>

Lianmin Zheng's avatar
Lianmin Zheng committed
12
13
--------------------------------------------------------------------------------

14
| [**Blog**](https://lmsys.org/blog/2024-07-25-sglang-llama3/) | [**Paper**](https://arxiv.org/abs/2312.07104) | [**Join Slack**](https://join.slack.com/t/sgl-fru7574/shared_invite/zt-2ngly9muu-t37XiH87qvD~6rVBTkTEHw) | [**Join Weekly Development Meeting**](https://t.co/4BFjCLnVHq) |
Lianmin Zheng's avatar
Lianmin Zheng committed
15

Ying Sheng's avatar
Ying Sheng committed
16
17
SGLang is a fast serving framework for large language models and vision language models.
It makes your interaction with models faster and more controllable by co-designing the backend runtime and frontend language.
18
The core features include:
19
20
21
22

- **Fast Backend Runtime**: Provides efficient serving with RadixAttention for prefix caching, jump-forward constrained decoding, continuous batching, token attention (paged attention), tensor parallelism, FlashInfer kernels, chunked prefill, and quantization (INT4/FP8/AWQ/GPTQ).
- **Flexible Frontend Language**: Offers an intuitive interface for programming LLM applications, including chained generation calls, advanced prompting, control flow, multi-modal inputs, parallelism, and external interactions.
- **Extensive Model Support**: Supports a wide range of generative models (Llama 3, Gemma 2, Mistral, QWen, DeepSeek, LLaVA, etc.) and embedding models (e5-mistral), with easy extensibility for integrating new models.
23
- **Active Community**: SGLang is open-source and backed by an active community with industry adoption.
Lianmin Zheng's avatar
Lianmin Zheng committed
24

Ying Sheng's avatar
Ying Sheng committed
25
## News
Yineng Zhang's avatar
Yineng Zhang committed
26
- [2024/09] 🔥 SGLang v0.3 Release: 7x Faster DeepSeek MLA, 1.5x Faster torch.compile, Multi-Image/Video LLaVA-OneVision ([blog](https://lmsys.org/blog/2024-09-04-sglang-v0-3/)).
Ying Sheng's avatar
Ying Sheng committed
27
28
- [2024/07] 🔥 Faster Llama3 Serving with SGLang Runtime (vs. TensorRT-LLM, vLLM) ([blog](https://lmsys.org/blog/2024-07-25-sglang-llama3/)).
- [2024/02] SGLang enables **3x faster JSON decoding** with compressed finite state machine ([blog](https://lmsys.org/blog/2024-02-05-compressed-fsm/)).
Ying Sheng's avatar
Ying Sheng committed
29

Ying Sheng's avatar
Ying Sheng committed
30
31
32
<details>
<summary>More</summary>

33
- [2024/04] SGLang is used by the official **LLaVA-NeXT (video)** release ([blog](https://llava-vl.github.io/blog/2024-04-30-llava-next-video/)).
Ying Sheng's avatar
Ying Sheng committed
34
- [2024/01] SGLang provides up to **5x faster inference** with RadixAttention ([blog](https://lmsys.org/blog/2024-01-17-sglang/)).
Ying Sheng's avatar
Ying Sheng committed
35
36
37
38
- [2024/01] SGLang powers the serving of the official **LLaVA v1.6** release demo ([usage](https://github.com/haotian-liu/LLaVA?tab=readme-ov-file#demo)).

</details>

Lianmin Zheng's avatar
Lianmin Zheng committed
39
40
41
## Contents
- [Install](#install)
- [Backend: SGLang Runtime (SRT)](#backend-sglang-runtime-srt)
Ying Sheng's avatar
Ying Sheng committed
42
- [Frontend: Structured Generation Language (SGLang)](#frontend-structured-generation-language-sglang)
Lianmin Zheng's avatar
Lianmin Zheng committed
43
44
45
46
47
48
- [Benchmark And Performance](#benchmark-and-performance)
- [Roadmap](#roadmap)
- [Citation And Acknowledgment](#citation-and-acknowledgment)

## Install

49
50
You can install SGLang using any of the methods below.

Lianmin Zheng's avatar
Lianmin Zheng committed
51
52
### Method 1: With pip
```
53
pip install --upgrade pip
Lianmin Zheng's avatar
Lianmin Zheng committed
54
pip install "sglang[all]"
Lianmin Zheng's avatar
Lianmin Zheng committed
55

Lianmin Zheng's avatar
Lianmin Zheng committed
56
# Install FlashInfer CUDA kernels
57
pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/
Lianmin Zheng's avatar
Lianmin Zheng committed
58
```
59

Lianmin Zheng's avatar
Lianmin Zheng committed
60
### Method 2: From source
Lianmin Zheng's avatar
Lianmin Zheng committed
61
```
Ying Sheng's avatar
Ying Sheng committed
62
# Use the last release branch
Ying Sheng's avatar
Ying Sheng committed
63
git clone -b v0.3.2 https://github.com/sgl-project/sglang.git
Lianmin Zheng's avatar
Lianmin Zheng committed
64
65
cd sglang

Lianmin Zheng's avatar
Lianmin Zheng committed
66
pip install --upgrade pip
Lianmin Zheng's avatar
Lianmin Zheng committed
67
68
pip install -e "python[all]"

Lianmin Zheng's avatar
Lianmin Zheng committed
69
# Install FlashInfer CUDA kernels
70
pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/
Lianmin Zheng's avatar
Lianmin Zheng committed
71
```
72

Lianmin Zheng's avatar
Lianmin Zheng committed
73
### Method 3: Using docker
74
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).
75
Replace `<secret>` below with your huggingface hub [token](https://huggingface.co/docs/hub/en/security-tokens).
Ying Sheng's avatar
Ying Sheng committed
76

Liangsheng Yin's avatar
Liangsheng Yin committed
77
78
79
80
```bash
docker run --gpus all \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
81
    --env "HF_TOKEN=<secret>" \
Liangsheng Yin's avatar
Liangsheng Yin committed
82
83
    --ipc=host \
    lmsysorg/sglang:latest \
84
    python3 -m sglang.launch_server --model-path meta-llama/Meta-Llama-3.1-8B-Instruct --host 0.0.0.0 --port 30000
Liangsheng Yin's avatar
Liangsheng Yin committed
85
86
```

87
88
### Method 4: Using docker compose

89
<details>
Dr. Artificial曾小健's avatar
Dr. Artificial曾小健 committed
90
<summary>More</summary>
91

92
> This method is recommended if you plan to serve it as a service.
93
> A better approach is to use the [k8s-sglang-service.yaml](docker/k8s-sglang-service.yaml).
94

95
1. Copy the [compose.yml](docker/compose.yaml) to your local machine
96
2. Execute the command `docker compose up -d` in your terminal.
97
</details>
98

99
100
### Method 5: Run on Kubernetes or Clouds with SkyPilot

101
<details>
Dr. Artificial曾小健's avatar
Dr. Artificial曾小健 committed
102
<summary>More</summary>
103

104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
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:
<details>
<summary>SkyPilot YAML: <code>sglang.yaml</code></summary>

```yaml
# sglang.yaml
envs:
  HF_TOKEN: null

resources:
  image_id: docker:lmsysorg/sglang:latest
  accelerators: A100
  ports: 30000

run: |
  conda deactivate
  python3 -m sglang.launch_server \
    --model-path meta-llama/Meta-Llama-3.1-8B-Instruct \
    --host 0.0.0.0 \
    --port 30000
```
128
</details>
129
130
131
132
133
134
135
136
137

```bash
# Deploy on any cloud or Kubernetes cluster. Use --cloud <cloud> to select a specific cloud provider.
HF_TOKEN=<secret> sky launch -c sglang --env HF_TOKEN sglang.yaml

# Get the HTTP API endpoint
sky status --endpoint 30000 sglang
```
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).
138
</details>
139
140


Lianmin Zheng's avatar
Lianmin Zheng committed
141
### Common Notes
Lianmin Zheng's avatar
Lianmin Zheng committed
142
- [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.
Lianmin Zheng's avatar
Lianmin Zheng committed
143
- If you only need to use the OpenAI backend, you can avoid installing other dependencies by using `pip install "sglang[openai]"`.
Ying Sheng's avatar
Ying Sheng committed
144

Ying Sheng's avatar
Ying Sheng committed
145
146
147
## Backend: SGLang Runtime (SRT)
The SGLang Runtime (SRT) is an efficient serving engine.

Ying Sheng's avatar
Ying Sheng committed
148
### Quick Start
Ying Sheng's avatar
Ying Sheng committed
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
Launch a server
```
python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000
```

Send a request
```
curl http://localhost:30000/generate \
  -H "Content-Type: application/json" \
  -d '{
    "text": "Once upon a time,",
    "sampling_params": {
      "max_new_tokens": 16,
      "temperature": 0
    }
  }'
```
166
Learn more about the argument format [here](docs/en/sampling_params.md).
Ying Sheng's avatar
Ying Sheng committed
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195

### OpenAI Compatible API
In addition, the server supports OpenAI-compatible APIs.

```python
import openai
client = openai.Client(
    base_url="http://127.0.0.1:30000/v1", api_key="EMPTY")

# Text completion
response = client.completions.create(
	model="default",
	prompt="The capital of France is",
	temperature=0,
	max_tokens=32,
)
print(response)

# Chat completion
response = client.chat.completions.create(
    model="default",
    messages=[
        {"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)
196
197
198
199
200
201
202

# Text embedding
response = client.embeddings.create(
    model="default",
    input="How are you today",
)
print(response)
Ying Sheng's avatar
Ying Sheng committed
203
204
```

Ying Sheng's avatar
Ying Sheng committed
205
It supports streaming, vision, and most features of the Chat/Completions/Models/Batch endpoints specified by the [OpenAI API Reference](https://platform.openai.com/docs/api-reference/).
Ying Sheng's avatar
Ying Sheng committed
206
207

### Additional Server Arguments
208
- 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.
Ying Sheng's avatar
Ying Sheng committed
209
```
210
python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --tp 2
Ying Sheng's avatar
Ying Sheng committed
211
```
212
- 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.
Ying Sheng's avatar
Ying Sheng committed
213
```
214
python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --dp 2 --tp 2
Ying Sheng's avatar
Ying Sheng committed
215
```
216
- 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`.
Ying Sheng's avatar
Ying Sheng committed
217
```
218
python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --mem-fraction-static 0.7
Ying Sheng's avatar
Ying Sheng committed
219
```
220
221
- See [hyperparameter_tuning.md](docs/en/hyperparameter_tuning.md) on tuning hyperparameters for better performance.
- If you see out-of-memory errors during prefill for long prompts, try to set a smaller chunked prefill size.
Ying Sheng's avatar
Ying Sheng committed
222
```
223
python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --chunked-prefill-size 4096
Ying Sheng's avatar
Ying Sheng committed
224
```
225
226
227
228
229
- To enable torch.compile acceleration, add `--enable-torch-compile`. It accelerates small models on small batch sizes.
- 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/en/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.
Ying Sheng's avatar
Ying Sheng committed
230
231
```
# Node 0
232
python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --tp 4 --nccl-init sgl-dev-0:50000 --nnodes 2 --node-rank 0
Ying Sheng's avatar
Ying Sheng committed
233
234

# Node 1
235
python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --tp 4 --nccl-init sgl-dev-0:50000 --nnodes 2 --node-rank 1
Ying Sheng's avatar
Ying Sheng committed
236
```
Lianmin Zheng's avatar
Lianmin Zheng committed
237
 
Ying Sheng's avatar
Ying Sheng committed
238
239
### Supported Models

240
**Generative Models**
241
- Llama / Llama 2 / Llama 3 / Llama 3.1
242
- Mistral / Mixtral / Mistral NeMo
Ying Sheng's avatar
Ying Sheng committed
243
- Gemma / Gemma 2
Niklas Muennighoff's avatar
Niklas Muennighoff committed
244
- OLMoE
Ying Sheng's avatar
Ying Sheng committed
245
- Qwen / Qwen 2 / Qwen 2 MoE
246
- DeepSeek / DeepSeek 2
247
- [LLaVA-OneVision](https://llava-vl.github.io/blog/2024-08-05-llava-onevision/)
248
  - `python3 -m sglang.launch_server --model-path lmms-lab/llava-onevision-qwen2-7b-ov --port=30000 --chat-template=chatml-llava`
249
  - `python3 -m sglang.launch_server --model-path lmms-lab/llava-onevision-qwen2-72b-ov --port=30000 --tp-size=8 --chat-template=chatml-llava`
250
251
252
253
254
  - 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)
- LLaVA 1.5 / 1.6 / NeXT
  - `python -m sglang.launch_server --model-path lmms-lab/llama3-llava-next-8b --port=30000 --tp-size=1 --chat-template=llava_llama_3`
  - `python -m sglang.launch_server --model-path lmms-lab/llava-next-72b --port=30000 --tp-size=8 --chat-template=chatml-llava`
  - 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)
Ying Sheng's avatar
Ying Sheng committed
255
256
257
258
259
260
261
- Yi-VL
- StableLM
- Command-R
- DBRX
- Grok
- ChatGLM
- InternLM 2
262
- Exaone 3
Vectory's avatar
Vectory committed
263
- BaiChuan2
William's avatar
William committed
264
- MiniCPM / MiniCPM 3
265
- XVERSE / XVERSE MoE
266
- SmolLM
William's avatar
William committed
267
268


269
270
271
272
273
274
**Embedding Models**

- e5-mistral
- gte-Qwen2
  - `python -m sglang.launch_server --model-path Alibaba-NLP/gte-Qwen2-7B-instruct --is-embedding`

275
Instructions for supporting a new model are [here](docs/en/model_support.md).
Ying Sheng's avatar
Ying Sheng committed
276

Lianmin Zheng's avatar
Lianmin Zheng committed
277
#### Use Models From ModelScope
278
<details>
Dr. Artificial曾小健's avatar
Dr. Artificial曾小健 committed
279
<summary>More</summary>
280

Lianmin Zheng's avatar
Lianmin Zheng committed
281
To use a model from [ModelScope](https://www.modelscope.cn), set the environment variable SGLANG_USE_MODELSCOPE.
Lianmin Zheng's avatar
Lianmin Zheng committed
282
283
284
285
286
287
```
export SGLANG_USE_MODELSCOPE=true
```
Launch [Qwen2-7B-Instruct](https://www.modelscope.cn/models/qwen/qwen2-7b-instruct) Server
```
SGLANG_USE_MODELSCOPE=true python -m sglang.launch_server --model-path qwen/Qwen2-7B-Instruct --port 30000
Lianmin Zheng's avatar
Lianmin Zheng committed
288
```
289
290
291
292
293
294
295
296
297
298
299

Or start it by docker.
```bash
docker run --gpus all \
    -p 30000:30000 \
    -v ~/.cache/modelscope:/root/.cache/modelscope \
    --env "SGLANG_USE_MODELSCOPE=true" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server --model-path Qwen/Qwen2.5-7B-Instruct --host 0.0.0.0 --port 30000
```
Lianmin Zheng's avatar
Lianmin Zheng committed
300
  
301
</details>
Lianmin Zheng's avatar
Lianmin Zheng committed
302
303

#### Run Llama 3.1 405B
Lianmin Zheng's avatar
Lianmin Zheng committed
304
<details>
Dr. Artificial曾小健's avatar
Dr. Artificial曾小健 committed
305
<summary>More</summary>
Ying Sheng's avatar
Ying Sheng committed
306
307

```bash
308
# Run 405B (fp8) on a single node
Ying Sheng's avatar
Ying Sheng committed
309
310
python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3.1-405B-Instruct-FP8 --tp 8

311
312
313
# Run 405B (fp16) on two nodes
## on the first node, replace the `172.16.4.52:20000` with your own first node ip address and port
GLOO_SOCKET_IFNAME=eth0 python3 -m sglang.launch_server --model-path meta-llama/Meta-Llama-3.1-405B-Instruct --tp 16 --nccl-init-addr 172.16.4.52:20000 --nnodes 2 --node-rank 0 --disable-cuda-graph
Ying Sheng's avatar
Ying Sheng committed
314

315
316
## on the first node, replace the `172.16.4.52:20000` with your own first node ip address and port
GLOO_SOCKET_IFNAME=eth0 python3 -m sglang.launch_server --model-path meta-llama/Meta-Llama-3.1-405B-Instruct --tp 16 --nccl-init-addr 172.16.4.52:20000 --nnodes 2 --node-rank 1 --disable-cuda-graph
Ying Sheng's avatar
Ying Sheng committed
317
318
```

Lianmin Zheng's avatar
Lianmin Zheng committed
319
320
</details>

Ying Sheng's avatar
Ying Sheng committed
321
322
### Benchmark Performance

Lianmin Zheng's avatar
Lianmin Zheng committed
323
324
325
- 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.
Ying Sheng's avatar
Ying Sheng committed
326
  ```
327
  python -m sglang.bench_latency --model-path meta-llama/Meta-Llama-3-8B-Instruct --batch 32 --input-len 256 --output-len 32
Ying Sheng's avatar
Ying Sheng committed
328
329
330
  ```
- Benchmark online serving. Launch a server first and run the following command.
  ```
331
  python3 -m sglang.bench_serving --backend sglang --num-prompt 10
Ying Sheng's avatar
Ying Sheng committed
332
333
  ```

Ying Sheng's avatar
Ying Sheng committed
334
## Frontend: Structured Generation Language (SGLang)
335
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.
Ying Sheng's avatar
Ying Sheng committed
336
337

### Quick Start
Lianmin Zheng's avatar
Lianmin Zheng committed
338
339
The example below shows how to use sglang to answer a mulit-turn question.

Ying Sheng's avatar
Ying Sheng committed
340
#### Using Local Models
341
First, launch a server with
Lianmin Zheng's avatar
Lianmin Zheng committed
342
```
Ying Sheng's avatar
Ying Sheng committed
343
python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000
Lianmin Zheng's avatar
Lianmin Zheng committed
344
345
```

346
347
Then, connect to the server and answer a multi-turn question.

Lianmin Zheng's avatar
Lianmin Zheng committed
348
```python
349
from sglang import function, system, user, assistant, gen, set_default_backend, RuntimeEndpoint
Lianmin Zheng's avatar
Lianmin Zheng committed
350
351
352
353
354
355
356
357
358

@function
def multi_turn_question(s, question_1, question_2):
    s += system("You are a helpful assistant.")
    s += user(question_1)
    s += assistant(gen("answer_1", max_tokens=256))
    s += user(question_2)
    s += assistant(gen("answer_2", max_tokens=256))

359
set_default_backend(RuntimeEndpoint("http://localhost:30000"))
Lianmin Zheng's avatar
Lianmin Zheng committed
360
361
362
363
364
365
366
367

state = multi_turn_question.run(
    question_1="What is the capital of the United States?",
    question_2="List two local attractions.",
)

for m in state.messages():
    print(m["role"], ":", m["content"])
368
369

print(state["answer_1"])
Lianmin Zheng's avatar
Lianmin Zheng committed
370
371
```

Ying Sheng's avatar
Ying Sheng committed
372
#### Using OpenAI Models
373
Set the OpenAI API Key
Lianmin Zheng's avatar
Lianmin Zheng committed
374
```
375
export OPENAI_API_KEY=sk-******
Lianmin Zheng's avatar
Lianmin Zheng committed
376
377
```

378
Then, answer a multi-turn question.
Lianmin Zheng's avatar
Lianmin Zheng committed
379
```python
380
from sglang import function, system, user, assistant, gen, set_default_backend, OpenAI
Lianmin Zheng's avatar
Lianmin Zheng committed
381
382
383
384
385
386
387
388
389

@function
def multi_turn_question(s, question_1, question_2):
    s += system("You are a helpful assistant.")
    s += user(question_1)
    s += assistant(gen("answer_1", max_tokens=256))
    s += user(question_2)
    s += assistant(gen("answer_2", max_tokens=256))

390
set_default_backend(OpenAI("gpt-3.5-turbo"))
Lianmin Zheng's avatar
Lianmin Zheng committed
391
392
393
394
395
396
397
398

state = multi_turn_question.run(
    question_1="What is the capital of the United States?",
    question_2="List two local attractions.",
)

for m in state.messages():
    print(m["role"], ":", m["content"])
399
400

print(state["answer_1"])
Lianmin Zheng's avatar
Lianmin Zheng committed
401
402
```

Ying Sheng's avatar
Ying Sheng committed
403
#### More Examples
Lianmin Zheng's avatar
Lianmin Zheng committed
404

405
Anthropic and VertexAI (Gemini) models are also supported.
Byron Hsu's avatar
Byron Hsu committed
406
You can find more examples at [examples/quick_start](examples/frontend_language/quick_start).
Lianmin Zheng's avatar
Lianmin Zheng committed
407

Ying Sheng's avatar
Ying Sheng committed
408
### Language Feature
Lianmin Zheng's avatar
Lianmin Zheng committed
409
410
411
412
413
To begin with, import sglang.
```python
import sglang as sgl
```

Lianmin Zheng's avatar
Lianmin Zheng committed
414
`sglang` provides some simple primitives such as `gen`, `select`, `fork`, `image`.
Lianmin Zheng's avatar
Lianmin Zheng committed
415
416
You can implement your prompt flow in a function decorated by `sgl.function`.
You can then invoke the function with `run` or `run_batch`.
417
The system will manage the state, chat template, parallelism and batching for you.
Lianmin Zheng's avatar
Lianmin Zheng committed
418

419
The complete code for the examples below can be found at [readme_examples.py](examples/frontend_language/usage/readme_examples.py)
420

Ying Sheng's avatar
Ying Sheng committed
421
#### Control Flow
Lianmin Zheng's avatar
Lianmin Zheng committed
422
423
You can use any Python code within the function body, including control flow, nested function calls, and external libraries.

Lianmin Zheng's avatar
Lianmin Zheng committed
424
425
```python
@sgl.function
426
427
428
def tool_use(s, question):
    s += "To answer this question: " + question + ". "
    s += "I need to use a " + sgl.gen("tool", choices=["calculator", "search engine"]) + ". "
Lianmin Zheng's avatar
Lianmin Zheng committed
429
430
431

    if s["tool"] == "calculator":
        s += "The math expression is" + sgl.gen("expression")
432
433
    elif s["tool"] == "search engine":
        s += "The key word to search is" + sgl.gen("word")
Lianmin Zheng's avatar
Lianmin Zheng committed
434
```
Lianmin Zheng's avatar
Lianmin Zheng committed
435

Ying Sheng's avatar
Ying Sheng committed
436
#### Parallelism
Lianmin Zheng's avatar
Lianmin Zheng committed
437
438
439
Use `fork` to launch parallel prompts.
Because `sgl.gen` is non-blocking, the for loop below issues two generation calls in parallel.

Lianmin Zheng's avatar
Lianmin Zheng committed
440
441
442
443
444
445
446
447
```python
@sgl.function
def tip_suggestion(s):
    s += (
        "Here are two tips for staying healthy: "
        "1. Balanced Diet. 2. Regular Exercise.\n\n"
    )

Lianmin Zheng's avatar
Lianmin Zheng committed
448
    forks = s.fork(2)
Lianmin Zheng's avatar
Lianmin Zheng committed
449
450
451
452
453
454
455
456
    for i, f in enumerate(forks):
        f += f"Now, expand tip {i+1} into a paragraph:\n"
        f += sgl.gen(f"detailed_tip", max_tokens=256, stop="\n\n")

    s += "Tip 1:" + forks[0]["detailed_tip"] + "\n"
    s += "Tip 2:" + forks[1]["detailed_tip"] + "\n"
    s += "In summary" + sgl.gen("summary")
```
Lianmin Zheng's avatar
Lianmin Zheng committed
457

Lianmin Zheng's avatar
Lianmin Zheng committed
458
#### Multi-Modality
Lianmin Zheng's avatar
Lianmin Zheng committed
459
460
Use `sgl.image` to pass an image as input.

Lianmin Zheng's avatar
Lianmin Zheng committed
461
462
```python
@sgl.function
Lianmin Zheng's avatar
Lianmin Zheng committed
463
def image_qa(s, image_file, question):
Lianmin Zheng's avatar
Lianmin Zheng committed
464
    s += sgl.user(sgl.image(image_file) + question)
Lianmin Zheng's avatar
Lianmin Zheng committed
465
    s += sgl.assistant(sgl.gen("answer", max_tokens=256)
Lianmin Zheng's avatar
Lianmin Zheng committed
466
467
```

468
See also [srt_example_llava.py](examples/frontend_language/quick_start/local_example_llava_next.py).
469

Ying Sheng's avatar
Ying Sheng committed
470
#### Constrained Decoding
471
472
Use `regex` to specify a regular expression as a decoding constraint.
This is only supported for local models.
Lianmin Zheng's avatar
Lianmin Zheng committed
473

Lianmin Zheng's avatar
Lianmin Zheng committed
474
```python
Lianmin Zheng's avatar
Lianmin Zheng committed
475
@sgl.function
Lianmin Zheng's avatar
Lianmin Zheng committed
476
477
def regular_expression_gen(s):
    s += "Q: What is the IP address of the Google DNS servers?\n"
Lianmin Zheng's avatar
Lianmin Zheng committed
478
    s += "A: " + sgl.gen(
Lianmin Zheng's avatar
Lianmin Zheng committed
479
480
481
482
483
        "answer",
        temperature=0,
        regex=r"((25[0-5]|2[0-4]\d|[01]?\d\d?).){3}(25[0-5]|2[0-4]\d|[01]?\d\d?)",
    )
```
Lianmin Zheng's avatar
Lianmin Zheng committed
484

Ying Sheng's avatar
Ying Sheng committed
485
#### JSON Decoding
Lianmin Zheng's avatar
Lianmin Zheng committed
486
Use `regex` to specify a JSON schema with a regular expression.
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507

```python
character_regex = (
    r"""\{\n"""
    + r"""    "name": "[\w\d\s]{1,16}",\n"""
    + r"""    "house": "(Gryffindor|Slytherin|Ravenclaw|Hufflepuff)",\n"""
    + r"""    "blood status": "(Pure-blood|Half-blood|Muggle-born)",\n"""
    + r"""    "occupation": "(student|teacher|auror|ministry of magic|death eater|order of the phoenix)",\n"""
    + r"""    "wand": \{\n"""
    + r"""        "wood": "[\w\d\s]{1,16}",\n"""
    + r"""        "core": "[\w\d\s]{1,16}",\n"""
    + r"""        "length": [0-9]{1,2}\.[0-9]{0,2}\n"""
    + r"""    \},\n"""
    + r"""    "alive": "(Alive|Deceased)",\n"""
    + r"""    "patronus": "[\w\d\s]{1,16}",\n"""
    + r"""    "bogart": "[\w\d\s]{1,16}"\n"""
    + r"""\}"""
)

@sgl.function
def character_gen(s, name):
Lianmin Zheng's avatar
Lianmin Zheng committed
508
    s += name + " is a character in Harry Potter. Please fill in the following information about this character.\n"
509
510
511
    s += sgl.gen("json_output", max_tokens=256, regex=character_regex)
```

512
See also [json_decode.py](examples/frontend_language/usage/json_decode.py) for an additional example of specifying formats with Pydantic models.
513

Ying Sheng's avatar
Ying Sheng committed
514
#### Batching
Lianmin Zheng's avatar
Lianmin Zheng committed
515
516
Use `run_batch` to run a batch of requests with continuous batching.

Lianmin Zheng's avatar
Lianmin Zheng committed
517
518
519
520
521
522
523
524
525
526
527
528
```python
@sgl.function
def text_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?"},
    ],
Lianmin Zheng's avatar
Lianmin Zheng committed
529
    progress_bar=True
Lianmin Zheng's avatar
Lianmin Zheng committed
530
531
)
```
Lianmin Zheng's avatar
Lianmin Zheng committed
532

Ying Sheng's avatar
Ying Sheng committed
533
#### Streaming
Lianmin Zheng's avatar
Lianmin Zheng committed
534
535
Add `stream=True` to enable streaming.

Lianmin Zheng's avatar
Lianmin Zheng committed
536
537
538
539
540
541
```python
@sgl.function
def text_qa(s, question):
    s += "Q: " + question + "\n"
    s += "A:" + sgl.gen("answer", stop="\n")

542
state = text_qa.run(
Lianmin Zheng's avatar
Lianmin Zheng committed
543
    question="What is the capital of France?",
Lianmin Zheng's avatar
Lianmin Zheng committed
544
545
546
    temperature=0.1,
    stream=True
)
Lianmin Zheng's avatar
Lianmin Zheng committed
547

Lianmin Zheng's avatar
Lianmin Zheng committed
548
549
550
for out in state.text_iter():
    print(out, end="", flush=True)
```
Lianmin Zheng's avatar
Lianmin Zheng committed
551

552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
#### 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
def chat_example(s):
    s += sgl.system("You are a helpful assistant.")
    # Same as: s += s.system("You are a helpful assistant.")

    with s.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()
```

Ying Sheng's avatar
Ying Sheng committed
570
#### Tips and Implementation Details
571
572
- 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`.
Lianmin Zheng's avatar
Lianmin Zheng committed
573

Ying Sheng's avatar
Ying Sheng committed
574
575
576
## Benchmark And Performance
![8b_throughput](https://lmsys.org/images/blog/sglang_llama3/8b_throughput.svg)
![70b_fp8_throughput](https://lmsys.org/images/blog/sglang_llama3/70b_fp8_throughput.svg)
Lianmin Zheng's avatar
Lianmin Zheng committed
577

Ying Sheng's avatar
Ying Sheng committed
578
Learn more at this [blog](https://lmsys.org/blog/2024-07-25-sglang-llama3/).
Lianmin Zheng's avatar
Lianmin Zheng committed
579

Lianmin Zheng's avatar
Lianmin Zheng committed
580
## Roadmap
581
[Development Roadmap (2024 Q4)](https://github.com/sgl-project/sglang/issues/1487)
Lianmin Zheng's avatar
Lianmin Zheng committed
582
583

## Citation And Acknowledgment
Ying Sheng's avatar
Ying Sheng committed
584
585
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).