sglang 0.2.10__tar.gz → 0.2.12__tar.gz

{sglang-0.2.10/sglang.egg-info → sglang-0.2.12}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sglang
-Version: 0.2.10
+Version: 0.2.12
 Summary: SGLang is yet another fast serving framework for large language models and vision language models.
 License:                                  Apache License
                                    Version 2.0, January 2004
@@ -220,7 +220,6 @@ Requires-Dist: fastapi; extra == "srt"
 Requires-Dist: hf_transfer; extra == "srt"
 Requires-Dist: huggingface_hub; extra == "srt"
 Requires-Dist: interegular; extra == "srt"
-Requires-Dist: jsonlines; extra == "srt"
 Requires-Dist: packaging; extra == "srt"
 Requires-Dist: pillow; extra == "srt"
 Requires-Dist: psutil; extra == "srt"
@@ -230,7 +229,7 @@ Requires-Dist: torch; extra == "srt"
 Requires-Dist: uvicorn; extra == "srt"
 Requires-Dist: uvloop; extra == "srt"
 Requires-Dist: zmq; extra == "srt"
-Requires-Dist: vllm==0.5.3.post1; extra == "srt"
+Requires-Dist: vllm==0.5.4; extra == "srt"
 Requires-Dist: outlines>=0.0.44; extra == "srt"
 Provides-Extra: openai
 Requires-Dist: openai>=1.0; extra == "openai"
@@ -239,11 +238,18 @@ Provides-Extra: anthropic
 Requires-Dist: anthropic>=0.20.0; extra == "anthropic"
 Provides-Extra: litellm
 Requires-Dist: litellm>=1.0.0; extra == "litellm"
+Provides-Extra: test
+Requires-Dist: jsonlines; extra == "test"
+Requires-Dist: matplotlib; extra == "test"
+Requires-Dist: pandas; extra == "test"
 Provides-Extra: all
 Requires-Dist: sglang[srt]; extra == "all"
 Requires-Dist: sglang[openai]; extra == "all"
 Requires-Dist: sglang[anthropic]; extra == "all"
 Requires-Dist: sglang[litellm]; extra == "all"
+Provides-Extra: dev
+Requires-Dist: sglang[all]; extra == "dev"
+Requires-Dist: sglang[test]; extra == "dev"
 
 <div align="center">
 <img src="https://raw.githubusercontent.com/sgl-project/sglang/main/assets/logo.png" alt="logo" width="400"></img>
@@ -296,20 +302,20 @@ pip install --upgrade pip
 pip install "sglang[all]"
 
 # Install FlashInfer CUDA kernels
-pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.3/
+pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/
 ```
 
 ### Method 2: From source
 ```
 # Use the last release branch
-git clone -b v0.2.10 https://github.com/sgl-project/sglang.git
+git clone -b v0.2.12 https://github.com/sgl-project/sglang.git
 cd sglang
 
 pip install --upgrade pip
 pip install -e "python[all]"
 
 # Install FlashInfer CUDA kernels
-pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.3/
+pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/
 ```
 
 ### Method 3: Using docker
@@ -383,22 +389,26 @@ response = client.chat.completions.create(
 print(response)
 ```
 
-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/).
+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/).
 
 ### 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.
+- Add `--tp 2` to enable multi-GPU tensor parallelism. If it reports the error "peer access is not supported between these two devices", add `--enable-p2p-check` to the server launch command.
 ```
 python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000 --tp 2
 ```
-- 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.
+- Add `--dp 2` to enable multi-GPU data parallelism. It can also be used together with tensor parallelism. Data parallelism is better for throughput if there is enough memory.
 ```
 python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000 --dp 2 --tp 2
 ```
-- 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`
+- 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`.
 ```
 python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000 --mem-fraction-static 0.7
 ```
 - 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.
+```
+python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000 --chunked-prefill-size 4096
+```
 - 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-0` be the hostname of the first node and `50000` be an available port.
 ```
 # Node 0
@@ -408,29 +418,13 @@ python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct
 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
 ```
 - If the model does not have a template in the Hugging Face tokenizer, you can specify a [custom chat template](docs/en/custom_chat_template.md).
-- To enable fp8 quantization, you can add `--quantization fp8` on a fp16 checkpoint or directly load a fp8 checkpoint without specifying any arguments.
 - To enable experimental torch.compile support, you can add `--enable-torch-compile`. It accelerates small models on small batch sizes.
-
-### Run Llama 3.1 405B
-
-```bash
-## Run 405B (fp8) on a single node
-python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3.1-405B-Instruct-FP8 --tp 8
-
-## Run 405B (fp16) on two nodes
-# replace the `172.16.4.52:20000` with your own first node ip address and port, disable CUDA Graph temporarily
-
-# on the first node
-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 --mem-frac 0.75
-
-# on the second
-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 --mem-frac 0.75
-```
-
+- To enable fp8 quantization, you can add `--quantization fp8` on a fp16 checkpoint or directly load a fp8 checkpoint without specifying any arguments.
+ 
 ### Supported Models
 
 - Llama / Llama 2 / Llama 3 / Llama 3.1
-- Mistral / Mixtral
+- Mistral / Mixtral / Mistral NeMo
 - Gemma / Gemma 2
 - Qwen / Qwen 2 / Qwen 2 MoE
 - DeepSeek / DeepSeek 2
@@ -448,10 +442,35 @@ GLOO_SOCKET_IFNAME=eth0 python3 -m sglang.launch_server --model-path meta-llama/
 - Grok
 - ChatGLM
 - InternLM 2
-- Mistral NeMo
 
 Instructions for supporting a new model are [here](https://github.com/sgl-project/sglang/blob/main/docs/en/model_support.md).
 
+#### Use Models From ModelScope
+To use model from [ModelScope](https://www.modelscope.cn), setting environment variable SGLANG_USE_MODELSCOPE.
+```
+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
+```    
+
+#### Run Llama 3.1 405B
+
+```bash
+## Run 405B (fp8) on a single node
+python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3.1-405B-Instruct-FP8 --tp 8
+
+## Run 405B (fp16) on two nodes
+# replace the `172.16.4.52:20000` with your own first node ip address and port, disable CUDA Graph temporarily
+
+# on the first node
+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 --mem-frac 0.75
+
+# on the second
+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 --mem-frac 0.75
+```
+
 ### Benchmark Performance
 
 - 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, consider using `sglang.bench_serving`.
@@ -464,7 +483,7 @@ Instructions for supporting a new model are [here](https://github.com/sgl-projec
   ```
 
 ## Frontend: Structured Generation Language (SGLang)
-The frontend language can be used with local models or API models.
+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
 The example below shows how to use sglang to answer a mulit-turn question.

{sglang-0.2.10 → sglang-0.2.12}/README.md

@@ -49,20 +49,20 @@ pip install --upgrade pip
 pip install "sglang[all]"
 
 # Install FlashInfer CUDA kernels
-pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.3/
+pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/
 ```
 
 ### Method 2: From source
 ```
 # Use the last release branch
-git clone -b v0.2.10 https://github.com/sgl-project/sglang.git
+git clone -b v0.2.12 https://github.com/sgl-project/sglang.git
 cd sglang
 
 pip install --upgrade pip
 pip install -e "python[all]"
 
 # Install FlashInfer CUDA kernels
-pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.3/
+pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/
 ```
 
 ### Method 3: Using docker
@@ -136,22 +136,26 @@ response = client.chat.completions.create(
 print(response)
 ```
 
-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/).
+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/).
 
 ### 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.
+- Add `--tp 2` to enable multi-GPU tensor parallelism. If it reports the error "peer access is not supported between these two devices", add `--enable-p2p-check` to the server launch command.
 ```
 python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000 --tp 2
 ```
-- 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.
+- Add `--dp 2` to enable multi-GPU data parallelism. It can also be used together with tensor parallelism. Data parallelism is better for throughput if there is enough memory.
 ```
 python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000 --dp 2 --tp 2
 ```
-- 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`
+- 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`.
 ```
 python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000 --mem-fraction-static 0.7
 ```
 - 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.
+```
+python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --port 30000 --chunked-prefill-size 4096
+```
 - 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-0` be the hostname of the first node and `50000` be an available port.
 ```
 # Node 0
@@ -161,29 +165,13 @@ python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct
 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
 ```
 - If the model does not have a template in the Hugging Face tokenizer, you can specify a [custom chat template](docs/en/custom_chat_template.md).
-- To enable fp8 quantization, you can add `--quantization fp8` on a fp16 checkpoint or directly load a fp8 checkpoint without specifying any arguments.
 - To enable experimental torch.compile support, you can add `--enable-torch-compile`. It accelerates small models on small batch sizes.
-
-### Run Llama 3.1 405B
-
-```bash
-## Run 405B (fp8) on a single node
-python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3.1-405B-Instruct-FP8 --tp 8
-
-## Run 405B (fp16) on two nodes
-# replace the `172.16.4.52:20000` with your own first node ip address and port, disable CUDA Graph temporarily
-
-# on the first node
-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 --mem-frac 0.75
-
-# on the second
-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 --mem-frac 0.75
-```
-
+- To enable fp8 quantization, you can add `--quantization fp8` on a fp16 checkpoint or directly load a fp8 checkpoint without specifying any arguments.
+ 
 ### Supported Models
 
 - Llama / Llama 2 / Llama 3 / Llama 3.1
-- Mistral / Mixtral
+- Mistral / Mixtral / Mistral NeMo
 - Gemma / Gemma 2
 - Qwen / Qwen 2 / Qwen 2 MoE
 - DeepSeek / DeepSeek 2
@@ -201,10 +189,35 @@ GLOO_SOCKET_IFNAME=eth0 python3 -m sglang.launch_server --model-path meta-llama/
 - Grok
 - ChatGLM
 - InternLM 2
-- Mistral NeMo
 
 Instructions for supporting a new model are [here](https://github.com/sgl-project/sglang/blob/main/docs/en/model_support.md).
 
+#### Use Models From ModelScope
+To use model from [ModelScope](https://www.modelscope.cn), setting environment variable SGLANG_USE_MODELSCOPE.
+```
+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
+```    
+
+#### Run Llama 3.1 405B
+
+```bash
+## Run 405B (fp8) on a single node
+python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3.1-405B-Instruct-FP8 --tp 8
+
+## Run 405B (fp16) on two nodes
+# replace the `172.16.4.52:20000` with your own first node ip address and port, disable CUDA Graph temporarily
+
+# on the first node
+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 --mem-frac 0.75
+
+# on the second
+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 --mem-frac 0.75
+```
+
 ### Benchmark Performance
 
 - 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, consider using `sglang.bench_serving`.
@@ -217,7 +230,7 @@ Instructions for supporting a new model are [here](https://github.com/sgl-projec
   ```
 
 ## Frontend: Structured Generation Language (SGLang)
-The frontend language can be used with local models or API models.
+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
 The example below shows how to use sglang to answer a mulit-turn question.

{sglang-0.2.10 → sglang-0.2.12}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "sglang"
-version = "0.2.10"
+version = "0.2.12"
 description = "SGLang is yet another fast serving framework for large language models and vision language models."
 readme = "README.md"
 requires-python = ">=3.8"
@@ -20,14 +20,16 @@ dependencies = [
 ]
 
 [project.optional-dependencies]
-srt = ["aiohttp", "fastapi", "hf_transfer", "huggingface_hub", "interegular", "jsonlines",
+srt = ["aiohttp", "fastapi", "hf_transfer", "huggingface_hub", "interegular",
        "packaging", "pillow", "psutil", "pydantic", "python-multipart",
        "torch", "uvicorn", "uvloop", "zmq",
-       "vllm==0.5.3.post1", "outlines>=0.0.44"]
+       "vllm==0.5.4", "outlines>=0.0.44"]
 openai = ["openai>=1.0", "tiktoken"]
 anthropic = ["anthropic>=0.20.0"]
 litellm = ["litellm>=1.0.0"]
+test = ["jsonlines", "matplotlib", "pandas"]
 all = ["sglang[srt]", "sglang[openai]", "sglang[anthropic]", "sglang[litellm]"]
+dev = ["sglang[all]", "sglang[test]"]
 
 [project.urls]
 "Homepage" = "https://github.com/sgl-project/sglang"

{sglang-0.2.10 → sglang-0.2.12}/sglang/__init__.py

@@ -22,6 +22,11 @@ from sglang.api import (
     user_end,
     video,
 )
+from sglang.lang.choices import (
+    greedy_token_selection,
+    token_length_normalized,
+    unconditional_likelihood_normalized,
+)
 
 # SGLang DSL APIs
 __all__ = [
@@ -45,6 +50,9 @@ __all__ = [
     "user_begin",
     "user_end",
     "video",
+    "greedy_token_selection",
+    "token_length_normalized",
+    "unconditional_likelihood_normalized",
 ]
 
 # Global Configurations

{sglang-0.2.10 → sglang-0.2.12}/sglang/api.py

@@ -6,6 +6,7 @@ from typing import Callable, List, Optional, Union
 
 from sglang.global_config import global_config
 from sglang.lang.backend.base_backend import BaseBackend
+from sglang.lang.choices import ChoicesSamplingMethod, token_length_normalized
 from sglang.lang.ir import (
     SglExpr,
     SglExprList,
@@ -73,12 +74,18 @@ def gen(
     return_text_in_logprobs: Optional[bool] = None,
     dtype: Optional[type] = None,
     choices: Optional[List[str]] = None,
+    choices_method: Optional[ChoicesSamplingMethod] = None,
     regex: Optional[str] = None,
 ):
     """Call the model to generate. See the meaning of the arguments in docs/en/sampling_params.md"""
 
     if choices:
-        return SglSelect(name, choices, 0.0 if temperature is None else temperature)
+        return SglSelect(
+            name,
+            choices,
+            0.0 if temperature is None else temperature,
+            token_length_normalized if choices_method is None else choices_method,
+        )
 
     # check regex is valid
     if regex is not None:
@@ -186,9 +193,10 @@ def select(
     name: Optional[str] = None,
     choices: Optional[List[str]] = None,
     temperature: float = 0.0,
+    choices_method: ChoicesSamplingMethod = token_length_normalized,
 ):
     assert choices is not None
-    return SglSelect(name, choices, temperature)
+    return SglSelect(name, choices, temperature, choices_method)
 
 
 def _role_common(name: str, expr: Optional[SglExpr] = None):