Improve docs (#662)

README.md

@@ -10,7 +10,7 @@ SGLang is a fast serving framework for large language models and vision language
 It makes your interaction with models faster and more controllable by co-designing the backend runtime and frontend language.
 
 The core features include:
-- **Fast Backend Runtime**: Efficient serving with RadixAttention for prefix caching, continuous batching, token attention (paged attention), tensor parallelism, flashinfer kernels, jump-forward constrained decoding, and quantization (AWQ/FP8/GPTQ/Marlin).
+- **Fast Backend Runtime**: Efficient serving with RadixAttention for prefix caching, jump-forward constrained decoding, continuous batching, token attention (paged attention), tensor parallelism, flashinfer kernels, and quantization (AWQ/FP8/GPTQ/Marlin).
 - **Flexible Frontend Language**: Enables easy programming of LLM applications with chained generation calls, advanced prompting, control flow, multiple modalities, parallelism, and external interactions.
 
 ## News
@@ -129,7 +129,7 @@ response = client.chat.completions.create(
 print(response)
 ```
 
-It supports streaming and most features of the Chat/Completions/Models endpoints specified by the [OpenAI API Reference](https://platform.openai.com/docs/api-reference/).
+It supports streaming, vision, and most features of the Chat/Completions/Models endpoints specified by the [OpenAI API Reference](https://platform.openai.com/docs/api-reference/).
 
 ### Additional Server Arguments
 - Add `--tp 2` to enable tensor parallelism. If it indicates `peer access is not supported between these two devices`, add `--enable-p2p-check` option.

docs/sampling_params.md

@@ -8,23 +8,24 @@ The `/generate` endpoint accepts the following arguments in the JSON format.
 class GenerateReqInput:
     # The input prompt. It can be a single prompt or a batch of prompts.
     text: Union[List[str], str]
-    # The token ids for text; one can either specify text or input_ids
+    # The token ids for text; one can either specify text or input_ids.
     input_ids: Optional[Union[List[List[int]], List[int]]] = None
-    # The image input. It can be a file name.
+    # The image input. It can be a file name, a url, or base64 encoded string.
+    # See also python/sglang/srt/utils.py:load_image.
     image_data: Optional[Union[List[str], str]] = None
-    # The sampling_params
+    # The sampling_params.
     sampling_params: Union[List[Dict], Dict] = None
-    # The request id
+    # The request id.
     rid: Optional[Union[List[str], str]] = None
-    # Whether to return logprobs
+    # Whether to return logprobs.
     return_logprob: Optional[Union[List[bool], bool]] = None
-    # The start location of the prompt for return_logprob
+    # The start location of the prompt for return_logprob.
     logprob_start_len: Optional[Union[List[int], int]] = None
-    # The number of top logprobs to return
+    # The number of top logprobs to return.
     top_logprobs_num: Optional[Union[List[int], int]] = None
-    # Whether to detokenize tokens in logprobs
+    # Whether to detokenize tokens in logprobs.
     return_text_in_logprobs: bool = False
-    # Whether to stream output
+    # Whether to stream output.
     stream: bool = False
 ```
 
@@ -48,13 +49,19 @@ class SamplingParams:
     ) -> None:
 ```
 
+- `max_new_tokens`, `stop`, `temperature`, `top_p`, `top_k` are common sampling parameters.
+- `ignore_eos` means ignoring the EOS token and continue decoding, which is helpful for benchmarking purposes.
+- `regex` constrains the output to follow a given regular expression.
+
 ## Examples
 
 ### Normal
+Launch a server
 ```
-python -m sglang.launch_server --model-path meta-llama/Llama-2-7b-chat-hf --port 30000
+python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000
 ```
 
+Send a request
 ```python
 import requests
 
@@ -72,7 +79,7 @@ print(response.json())
 ```
 
 ### Streaming
-
+Send a request and stream the output
 ```python
 import requests, json
 
@@ -104,4 +111,32 @@ print("")
 
 ### Multi modal
 
-See [test_httpserver_llava.py](../test/srt/test_httpserver_llava.py).
+Launch a server
+```
+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
+```
+
+Download an image
+```
+curl -o example_image.png -L https://github.com/sgl-project/sglang/blob/main/test/lang/example_image.png?raw=true
+```
+
+```python
+import requests
+
+response = requests.post(
+    "http://localhost:30000/generate",
+    json={
+        "text": "A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions. USER: <image>\nDescribe this picture ASSISTANT:",
+        "image_data": "example_image.png",
+        "sampling_params": {
+            "temperature": 0,
+            "max_new_tokens": 32,
+        },
+    },
+)
+print(response.json())
+```
+
+The `image_data` can be a file name, a URL, or a base64 encoded string. See also `python/sglang/srt/utils.py:load_image`.
+Streaming is supported in a similar manner as [above](#streaming).

python/sglang/README.md

@@ -4,7 +4,8 @@
 - `srt`: The backend engine for running local models. (SRT = SGLang Runtime).
 - `test`: Test utilities.
 - `api.py`: Public API.
-- `bench_latency.py`: Benchmark utilities.
+- `bench_latency.py`: Benchmark a single static batch.
+- `bench_serving.py`: Benchmark online serving with dynamic requests.
 - `global_config.py`: The global configs and constants.
 - `launch_server.py`: The entry point of launching local server.
 - `utils.py`: Common utilities.

python/sglang/check_env.py

@@ -1,3 +1,5 @@
+"""Check environment configurations and dependency versions."""
+
 import importlib
 import os
 import resource

python/sglang/srt/managers/io_struct.py

@@ -13,25 +13,26 @@ from sglang.srt.sampling_params import SamplingParams
 
 @dataclass
 class GenerateReqInput:
-    # The input prompt
-    text: Optional[Union[List[str], str]] = None
-    # The token ids for text; one can either specify text or input_ids
+    # The input prompt. It can be a single prompt or a batch of prompts.
+    text: Union[List[str], str]
+    # The token ids for text; one can either specify text or input_ids.
     input_ids: Optional[Union[List[List[int]], List[int]]] = None
-    # The image input
+    # The image input. It can be a file name, a url, or base64 encoded string.
+    # See also python/sglang/srt/utils.py:load_image.
     image_data: Optional[Union[List[str], str]] = None
-    # The sampling_params
+    # The sampling_params.
     sampling_params: Union[List[Dict], Dict] = None
-    # The request id
+    # The request id.
     rid: Optional[Union[List[str], str]] = None
-    # Whether to return logprobs
+    # Whether to return logprobs.
     return_logprob: Optional[Union[List[bool], bool]] = None
-    # The start location of the prompt for return_logprob
+    # The start location of the prompt for return_logprob.
     logprob_start_len: Optional[Union[List[int], int]] = None
-    # The number of top logprobs to return
+    # The number of top logprobs to return.
     top_logprobs_num: Optional[Union[List[int], int]] = None
-    # Whether to detokenize tokens in logprobs
+    # Whether to detokenize tokens in logprobs.
     return_text_in_logprobs: bool = False
-    # Whether to stream output
+    # Whether to stream output.
     stream: bool = False
 
     def post_init(self):

python/sglang/srt/server.py

@@ -74,21 +74,6 @@ async def health() -> Response:
     return Response(status_code=200)
 
 
-def get_model_list():
-    """Available models."""
-    model_names = [tokenizer_manager.model_path]
-    return model_names
-
-
-@app.get("/v1/models")
-def available_models():
-    """Show available models."""
-    model_cards = []
-    for model_name in get_model_list():
-        model_cards.append(ModelCard(id=model_name, root=model_name))
-    return ModelList(data=model_cards)
-
-
 @app.get("/get_model_info")
 async def get_model_info():
     result = {
@@ -154,6 +139,16 @@ async def openai_v1_chat_completions(raw_request: Request):
     return await v1_chat_completions(tokenizer_manager, raw_request)
 
 
+@app.get("/v1/models")
+def available_models():
+    """Show available models."""
+    model_names = [tokenizer_manager.model_path]
+    model_cards = []
+    for model_name in model_names:
+        model_cards.append(ModelCard(id=model_name, root=model_name))
+    return ModelList(data=model_cards)
+
+
 def _set_global_server_args(server_args: ServerArgs):
     global global_server_args_dict
     global_server_args_dict = {