README.md 15.8 KB
Newer Older
Lianmin Zheng's avatar
Lianmin Zheng committed
1
2
3
4
5
6
<div align="center">
<img src="assets/logo.png" alt="logo" width="400"></img>
</div>

--------------------------------------------------------------------------------

7
| [**Blog**](https://lmsys.org/blog/2024-01-17-sglang/) | [**Paper**](https://arxiv.org/abs/2312.07104) |
Lianmin Zheng's avatar
Lianmin Zheng committed
8
9
10
11

SGLang is a structured generation language designed for large language models (LLMs).
It makes your interaction with LLMs faster and more controllable by co-designing the frontend language and the runtime system.

12
The core features include:
Lianmin Zheng's avatar
Lianmin Zheng committed
13
- **Flexible Frontend Language**: Enables easy programming of LLM applications with chained generation calls, advanced prompting, control flow, multiple modalities, parallelism, and external interactions.
Lianmin Zheng's avatar
Lianmin Zheng committed
14
- **High-Performance Backend Runtime**: Features RadixAttention for accelerating complex LLM programs by reusing the KV cache across multiple calls. It can also serve as a standalone inference engine with all common techniques implemented (e.g., continuous batching and tensor parallelism).
Lianmin Zheng's avatar
Lianmin Zheng committed
15

Ying Sheng's avatar
Ying Sheng committed
16
## News
Lianmin Zheng's avatar
Lianmin Zheng committed
17
- [2024/02] 🔥 SGLang enables **3x faster JSON decoding** with compressed finite state machine ([blog](https://lmsys.org/blog/2024-02-05-compressed-fsm/)).
18
- [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)).
Lianmin Zheng's avatar
Lianmin Zheng committed
19
- [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
20

Lianmin Zheng's avatar
Lianmin Zheng committed
21
22
23
## Contents
- [Install](#install)
- [Quick Start](#quick-start)
24
- [Frontend: Structured Generation Language (SGLang)](#frontend-structured-generation-language-sglang)
Lianmin Zheng's avatar
Lianmin Zheng committed
25
26
27
28
29
30
31
- [Backend: SGLang Runtime (SRT)](#backend-sglang-runtime-srt)
- [Benchmark And Performance](#benchmark-and-performance)
- [Roadmap](#roadmap)
- [Citation And Acknowledgment](#citation-and-acknowledgment)

## Install

Lianmin Zheng's avatar
Lianmin Zheng committed
32
33
34
### Method 1: With pip
```
pip install "sglang[all]"
Lianmin Zheng's avatar
Lianmin Zheng committed
35

Lianmin Zheng's avatar
Lianmin Zheng committed
36
37
38
# Install FlashInfer CUDA kernels
pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.3/
```
39

Lianmin Zheng's avatar
Lianmin Zheng committed
40
### Method 2: From source
Lianmin Zheng's avatar
Lianmin Zheng committed
41
```
42
git clone https://github.com/sgl-project/sglang.git
Lianmin Zheng's avatar
Lianmin Zheng committed
43
44
45
46
cd sglang

pip install -e "python[all]"

Lianmin Zheng's avatar
Lianmin Zheng committed
47
48
49
# Install FlashInfer CUDA kernels
pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.3/
```
50

Lianmin Zheng's avatar
Lianmin Zheng committed
51
52
### Method 3: Using docker
The docker images are available on Docker Hub as [lmsysorg/sglang](https://hub.docker.com/r/lmsysorg/sglang/tags).
Ying Sheng's avatar
Ying Sheng committed
53

Lianmin Zheng's avatar
Lianmin Zheng committed
54
### Common Notes
55
56
57
58
59
- If you see errors from the Triton compiler, please install the [Triton Nightly](https://triton-lang.org/main/getting-started/installation.html) by
```
pip uninstall -y triton triton-nightly
pip install -U --index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/Triton-Nightly/pypi/simple/ triton-nightly
```
Lianmin Zheng's avatar
Lianmin Zheng committed
60
61
- If you cannot install FlashInfer, check out its [installation](https://docs.flashinfer.ai/installation.html#) page. If you still cannot install it, you can use the slower Triton kernels by adding `--disable-flashinfer` when launching the server.
- 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
62

Lianmin Zheng's avatar
Lianmin Zheng committed
63
64
65
## Quick Start
The example below shows how to use sglang to answer a mulit-turn question.

66
67
### Using Local Models
First, launch a server with
Lianmin Zheng's avatar
Lianmin Zheng committed
68
```
69
python -m sglang.launch_server --model-path meta-llama/Llama-2-7b-chat-hf --port 30000
Lianmin Zheng's avatar
Lianmin Zheng committed
70
71
```

72
73
Then, connect to the server and answer a multi-turn question.

Lianmin Zheng's avatar
Lianmin Zheng committed
74
```python
75
from sglang import function, system, user, assistant, gen, set_default_backend, RuntimeEndpoint
Lianmin Zheng's avatar
Lianmin Zheng committed
76
77
78
79
80
81
82
83
84

@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))

85
set_default_backend(RuntimeEndpoint("http://localhost:30000"))
Lianmin Zheng's avatar
Lianmin Zheng committed
86
87
88
89
90
91
92
93

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"])
94
95

print(state["answer_1"])
Lianmin Zheng's avatar
Lianmin Zheng committed
96
97
```

98
99
### Using OpenAI Models
Set the OpenAI API Key
Lianmin Zheng's avatar
Lianmin Zheng committed
100
```
101
export OPENAI_API_KEY=sk-******
Lianmin Zheng's avatar
Lianmin Zheng committed
102
103
```

104
Then, answer a multi-turn question.
Lianmin Zheng's avatar
Lianmin Zheng committed
105
```python
106
from sglang import function, system, user, assistant, gen, set_default_backend, OpenAI
Lianmin Zheng's avatar
Lianmin Zheng committed
107
108
109
110
111
112
113
114
115

@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))

116
set_default_backend(OpenAI("gpt-3.5-turbo"))
Lianmin Zheng's avatar
Lianmin Zheng committed
117
118
119
120
121
122
123
124

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"])
125
126

print(state["answer_1"])
Lianmin Zheng's avatar
Lianmin Zheng committed
127
128
129
130
```

### More Examples

131
Anthropic and VertexAI (Gemini) models are also supported.
Lianmin Zheng's avatar
Lianmin Zheng committed
132
133
You can find more examples at [examples/quick_start](examples/quick_start).

134
## Frontend: Structured Generation Language (SGLang)
Lianmin Zheng's avatar
Lianmin Zheng committed
135

Lianmin Zheng's avatar
Lianmin Zheng committed
136
137
138
139
140
To begin with, import sglang.
```python
import sglang as sgl
```

Lianmin Zheng's avatar
Lianmin Zheng committed
141
`sglang` provides some simple primitives such as `gen`, `select`, `fork`, `image`.
Lianmin Zheng's avatar
Lianmin Zheng committed
142
143
You can implement your prompt flow in a function decorated by `sgl.function`.
You can then invoke the function with `run` or `run_batch`.
144
The system will manage the state, chat template, parallelism and batching for you.
Lianmin Zheng's avatar
Lianmin Zheng committed
145

146
147
The complete code for the examples below can be found at [readme_examples.py](examples/usage/readme_examples.py)

Lianmin Zheng's avatar
Lianmin Zheng committed
148
### Control Flow
Lianmin Zheng's avatar
Lianmin Zheng committed
149
150
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
151
152
```python
@sgl.function
153
154
155
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
156
157
158

    if s["tool"] == "calculator":
        s += "The math expression is" + sgl.gen("expression")
159
160
    elif s["tool"] == "search engine":
        s += "The key word to search is" + sgl.gen("word")
Lianmin Zheng's avatar
Lianmin Zheng committed
161
```
Lianmin Zheng's avatar
Lianmin Zheng committed
162
163

### Parallelism
Lianmin Zheng's avatar
Lianmin Zheng committed
164
165
166
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
167
168
169
170
171
172
173
174
```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
175
    forks = s.fork(2)
Lianmin Zheng's avatar
Lianmin Zheng committed
176
177
178
179
180
181
182
183
    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
184
185

### Multi Modality
Lianmin Zheng's avatar
Lianmin Zheng committed
186
187
Use `sgl.image` to pass an image as input.

Lianmin Zheng's avatar
Lianmin Zheng committed
188
189
```python
@sgl.function
Lianmin Zheng's avatar
Lianmin Zheng committed
190
def image_qa(s, image_file, question):
Lianmin Zheng's avatar
Lianmin Zheng committed
191
    s += sgl.user(sgl.image(image_file) + question)
Lianmin Zheng's avatar
Lianmin Zheng committed
192
    s += sgl.assistant(sgl.gen("answer", max_tokens=256)
Lianmin Zheng's avatar
Lianmin Zheng committed
193
194
```

195
196
See also [srt_example_llava.py](examples/quick_start/srt_example_llava.py).

Lianmin Zheng's avatar
Lianmin Zheng committed
197
### Constrained Decoding
198
199
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
200

Lianmin Zheng's avatar
Lianmin Zheng committed
201
```python
Lianmin Zheng's avatar
Lianmin Zheng committed
202
@sgl.function
Lianmin Zheng's avatar
Lianmin Zheng committed
203
204
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
205
    s += "A: " + sgl.gen(
Lianmin Zheng's avatar
Lianmin Zheng committed
206
207
208
209
210
        "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
211

212
### JSON Decoding
Lianmin Zheng's avatar
Lianmin Zheng committed
213
Use `regex` to specify a JSON schema with a regular expression.
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234

```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
235
    s += name + " is a character in Harry Potter. Please fill in the following information about this character.\n"
236
237
238
    s += sgl.gen("json_output", max_tokens=256, regex=character_regex)
```

Lianmin Zheng's avatar
Lianmin Zheng committed
239
See also [json_decode.py](examples/usage/json_decode.py) for an additional example on specifying formats with Pydantic models.
240
241


Lianmin Zheng's avatar
Lianmin Zheng committed
242
### Batching
Lianmin Zheng's avatar
Lianmin Zheng committed
243
244
Use `run_batch` to run a batch of requests with continuous batching.

Lianmin Zheng's avatar
Lianmin Zheng committed
245
246
247
248
249
250
251
252
253
254
255
256
```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
257
    progress_bar=True
Lianmin Zheng's avatar
Lianmin Zheng committed
258
259
)
```
Lianmin Zheng's avatar
Lianmin Zheng committed
260
261

### Streaming
Lianmin Zheng's avatar
Lianmin Zheng committed
262
263
Add `stream=True` to enable streaming.

Lianmin Zheng's avatar
Lianmin Zheng committed
264
265
266
267
268
269
```python
@sgl.function
def text_qa(s, question):
    s += "Q: " + question + "\n"
    s += "A:" + sgl.gen("answer", stop="\n")

270
state = text_qa.run(
Lianmin Zheng's avatar
Lianmin Zheng committed
271
    question="What is the capital of France?",
Lianmin Zheng's avatar
Lianmin Zheng committed
272
273
274
    temperature=0.1,
    stream=True
)
Lianmin Zheng's avatar
Lianmin Zheng committed
275

Lianmin Zheng's avatar
Lianmin Zheng committed
276
277
278
for out in state.text_iter():
    print(out, end="", flush=True)
```
Lianmin Zheng's avatar
Lianmin Zheng committed
279

Lianmin Zheng's avatar
Lianmin Zheng committed
280
### Tips and Implementation Details
281
282
- 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
283

Lianmin Zheng's avatar
Lianmin Zheng committed
284
285
286
## Backend: SGLang Runtime (SRT)
The SGLang Runtime (SRT) is designed to work best with the SGLang frontend.
However, it can also be used as a standalone API server.
Ying Sheng's avatar
Ying Sheng committed
287
In this case, the [RadixAttention](https://arxiv.org/abs/2312.07104) can still greatly accelerate many use cases with automatic KV cache reuse.
Lianmin Zheng's avatar
Lianmin Zheng committed
288
289
290
291
292
293
294
295
296

### Usage
Launch a server
```
python -m sglang.launch_server --model-path meta-llama/Llama-2-7b-chat-hf --port 30000
```

Send a request
```
297
curl http://localhost:30000/generate \
Lianmin Zheng's avatar
Lianmin Zheng committed
298
299
  -H "Content-Type: application/json" \
  -d '{
300
    "text": "Once upon a time,",
301
    "sampling_params": {
302
303
304
      "max_new_tokens": 16,
      "temperature": 0
    }
Lianmin Zheng's avatar
Lianmin Zheng committed
305
306
  }'
```
307
308
Learn more about the argument format [here](docs/sampling_params.md).

309
310
311
312
313
314
315
### OpenAI Compatible API
In addition, the server supports an experimental OpenAI-compatible API.

```python
import openai
client = openai.Client(
    base_url="http://127.0.0.1:30000/v1", api_key="EMPTY")
Cody Yu's avatar
Cody Yu committed
316
317

# Text completion
318
319
320
321
322
323
324
response = client.completions.create(
	model="default",
	prompt="The capital of France is",
	temperature=0,
	max_tokens=32,
)
print(response)
Cody Yu's avatar
Cody Yu committed
325
326
327
328
329
330
331
332
333
334
335
336
337
338

# 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)
```

Lianmin Zheng's avatar
Lianmin Zheng committed
339
340
341
By default, the server uses the chat template specified in the model tokenizer from Hugging Face. It should just work for most official models such as Llama-2/Llama-3.

If needed, you can also override the chat template when launching the server:
Cody Yu's avatar
Cody Yu committed
342
343

```
344
python -m sglang.launch_server --model-path meta-llama/Llama-2-7b-chat-hf --port 30000 --chat-template llama-2
Cody Yu's avatar
Cody Yu committed
345
346
347
```

If the chat template you are looking for is missing, you are welcome to contribute it.
Lianmin Zheng's avatar
Lianmin Zheng committed
348
Meanwhile, you can also temporarily register your chat template as follows:
Cody Yu's avatar
Cody Yu committed
349
350
351
352
353
354
355
356
357
358
359
360
361
362

```json
{
  "name": "my_model",
  "system": "<|im_start|>system",
  "user": "<|im_start|>user",
  "assistant": "<|im_start|>assistant",
  "sep_style": "CHATML",
  "sep": "<|im_end|>",
  "stop_str": ["<|im_end|>", "<|im_start|>"]
}
```

```
363
python -m sglang.launch_server --model-path meta-llama/Llama-2-7b-chat-hf --port 30000 --chat-template ./my_model_template.json
364
365
```

Lianmin Zheng's avatar
Lianmin Zheng committed
366
### Additional Arguments
367
- Add `--tp 2` to enable tensor parallelism. If it indicates `peer access is not supported between these two devices`, add `--enable-p2p-check` option.
Lianmin Zheng's avatar
Lianmin Zheng committed
368
369
370
```
python -m sglang.launch_server --model-path meta-llama/Llama-2-7b-chat-hf --port 30000 --tp 2
```
Lianmin Zheng's avatar
Lianmin Zheng committed
371
372
373
374
- Add `--dp 2` to enable data parallelism. It can also be used together with tp. Data parallelism is better for throughput if there is enough memory.
```
python -m sglang.launch_server --model-path meta-llama/Llama-2-7b-chat-hf --port 30000 --dp 2 --tp 2
```
Ying Sheng's avatar
Ying Sheng committed
375
376
377
378
- If you see out-of-memory errors during serving, please 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`
```
python -m sglang.launch_server --model-path meta-llama/Llama-2-7b-chat-hf --port 30000 --mem-fraction-static 0.7
```
Lianmin Zheng's avatar
Lianmin Zheng committed
379
- See [hyperparameter_tuning.md](docs/hyperparameter_tuning.md) on tuning hyperparameters for better performance.
380
381
382
383
384
385
386
387
- Add `--nnodes 2` to run tensor parallelism on multiple nodes. If you have two nodes with two GPUs on each node and want to run TP=4, let `sgl-dev-1` be the hostname of the first node and `50000` be an available port.
```
# Node 0
python -m sglang.launch_server --model-path meta-llama/Llama-2-7b-chat-hf --tp 4 --nccl-init sgl-dev-1:50000 --nnodes 2 --node-rank 0

# Node 1
python -m sglang.launch_server --model-path meta-llama/Llama-2-7b-chat-hf --tp 4 --nccl-init sgl-dev-1:50000 --nnodes 2 --node-rank 1
```
Lianmin Zheng's avatar
Lianmin Zheng committed
388
389
390
391
392

### Supported Models
- Llama
- Mistral
- Mixtral
393
394
- Qwen / Qwen 2 / Qwen 2 MoE
- Gemma / Gemma 2
395
  - `python -m sglang.launch_server --model-path google/gemma-7b-it --port 30000 --attention-reduce-in-fp32`
Lianmin Zheng's avatar
Lianmin Zheng committed
396
- LLaVA
397
  - `python3 -m sglang.launch_server --model-path liuhaotian/llava-v1.5-7b --tokenizer-path llava-hf/llava-1.5-7b-hf --chat-template vicuna_v1.1 --port 30000`
398
399
  - `python3 -m sglang.launch_server --model-path liuhaotian/llava-v1.6-vicuna-7b --tokenizer-path llava-hf/llava-1.5-7b-hf --chat-template vicuna_v1.1 --port 30000`
  - `python3 -m sglang.launch_server --model-path liuhaotian/llava-v1.6-34b --tokenizer-path liuhaotian/llava-v1.6-34b-tokenizer --port 3000`
400
401
- LLaVA-NeXT-Video
  - see [srt_example_llava_v.sh](examples/usage/llava_video/srt_example_llava_v.sh)
Lianmin Zheng's avatar
Lianmin Zheng committed
402
403
- Yi-VL
  - see [srt_example_yi_vl.py](examples/quick_start/srt_example_yi_vl.py).
404
405
406
- StableLM
- Command-R
- DBRX
407
408
- Grok
- ChatGLM
409
410
411
- AWQ/GPTQ/Marlin quantization

Instructions for supporting a new model are [here](https://github.com/sgl-project/sglang/blob/main/docs/model_support.md).
Lianmin Zheng's avatar
Lianmin Zheng committed
412
413

## Benchmark And Performance
Lianmin Zheng's avatar
Lianmin Zheng committed
414
415
416
417
418
419
- Llama-7B on NVIDIA A10G, FP16, Tensor Parallelism=1
![llama_7b](assets/llama_7b.jpg)

- Mixtral-8x7B on NVIDIA A10G, FP16, Tensor Parallelism=8
![mixtral_8x7b](assets/mixtral_8x7b.jpg)

Lianmin Zheng's avatar
Lianmin Zheng committed
420
421
- Learn more about the above [results](docs/benchmark_results.md).
- Synthetic latency and throughput benchmark [scripts](https://github.com/sgl-project/sglang/tree/main/benchmark/latency_throughput).
Lianmin Zheng's avatar
Lianmin Zheng committed
422

Lianmin Zheng's avatar
Lianmin Zheng committed
423
## Roadmap
Ying Sheng's avatar
Ying Sheng committed
424
https://github.com/sgl-project/sglang/issues/157
Lianmin Zheng's avatar
Lianmin Zheng committed
425
426
427

## Citation And Acknowledgment
```
Lianmin Zheng's avatar
Lianmin Zheng committed
428
429
430
431
@misc{zheng2024sglang,
      title={SGLang: Efficient Execution of Structured Language Model Programs},
      author={Lianmin Zheng and Liangsheng Yin and Zhiqiang Xie and Chuyue Sun and Jeff Huang and Cody Hao Yu and Shiyi Cao and Christos Kozyrakis and Ion Stoica and Joseph E. Gonzalez and Clark Barrett and Ying Sheng},
      year={2024},
Lianmin Zheng's avatar
Lianmin Zheng committed
432
433
434
435
436
437
      eprint={2312.07104},
      archivePrefix={arXiv},
      primaryClass={cs.AI}
}
```

438
We learned from the design and reused some code of 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), [LMQL](https://github.com/eth-sri/lmql).