mirror of
https://git.datalinker.icu/vllm-project/vllm.git
synced 2026-01-10 00:30:59 +08:00
eace8bf0b9
This PR is the first step towards fixing https://github.com/vllm-project/vllm/pull/3208 It implements dynamic per-tensor scaling (see https://github.com/vllm-project/vllm/pull/4118), so users do not need to compute activation scales on a calibration dataset and they also don't need to convert their model checkpoints. It is enough to specify the `quantization="fp8"` argument. You can try out the PR like this: ```python from vllm import LLM, SamplingParams prompts = [ "Hello, my name is", "The president of the United States is", "The capital of France is", "The future of AI is", ] sampling_params = SamplingParams(temperature=0.8, top_p=0.95) llm = LLM(model="mistralai/Mixtral-8x7B-Instruct-v0.1", tensor_parallel_size=2, quantization="fp8") outputs = llm.generate(prompts, sampling_params) # Print the outputs. for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ``` **Performance**: For this PR, the focus is on making the code clean (while still trying to get reasonable performance), there is a bunch of optimizations that we will submit as a follow up PR that significantly improve the performance (similar to the numbers in https://github.com/vllm-project/vllm/pull/3954). With this PR, the results are as follows: <img width="725" alt="Screenshot 2024-04-21 at 1 31 50 PM" src="https://github.com/vllm-project/vllm/assets/113316/d8fe1118-07a0-4d4e-8530-37a77d465a03"> **Accuracy**: The accuracy with this PR on MMLU on `mistralai/Mixtral-8x7B-v0.1` is as follows: ``` | Groups |Version|Filter|n-shot|Metric|Value | |Stderr| |------------------|-------|------|-----:|------|-----:|---|-----:| |mmlu |N/A |none | 0|acc |0.7018|± |0.0036| | - humanities |N/A |none | 5|acc |0.6472|± |0.0065| | - other |N/A |none | 5|acc |0.7673|± |0.0072| | - social_sciences|N/A |none | 5|acc |0.8099|± |0.0070| | - stem |N/A |none | 5|acc |0.6131|± |0.0083| ``` this compares favorably with the fp16 results which are ``` | Groups |Version|Filter|n-shot|Metric|Value | |Stderr| |------------------|-------|------|-----:|------|-----:|---|-----:| |mmlu |N/A |none | 0|acc |0.7020|± |0.1313| | - humanities |N/A |none | 5|acc |0.6425|± |0.1349| | - other |N/A |none | 5|acc |0.7744|± |0.1038| | - social_sciences|N/A |none | 5|acc |0.8131|± |0.0695| | - stem |N/A |none | 5|acc |0.6108|± |0.1383| ``` Happy hacking!
Easy, fast, and cheap LLM serving for everyone
| Documentation | Blog | Paper | Discord |
Latest News 🔥
- [2024/04] We hosted the third vLLM meetup with Roblox! Please find the meetup slides here.
- [2024/01] We hosted the second vLLM meetup in SF! Please find the meetup slides here.
- [2024/01] Added ROCm 6.0 support to vLLM.
- [2023/12] Added ROCm 5.7 support to vLLM.
- [2023/10] We hosted the first vLLM meetup in SF! Please find the meetup slides here.
- [2023/09] We created our Discord server! Join us to discuss vLLM and LLM serving! We will also post the latest announcements and updates there.
- [2023/09] We released our PagedAttention paper on arXiv!
- [2023/08] We would like to express our sincere gratitude to Andreessen Horowitz (a16z) for providing a generous grant to support the open-source development and research of vLLM.
- [2023/07] Added support for LLaMA-2! You can run and serve 7B/13B/70B LLaMA-2s on vLLM with a single command!
- [2023/06] Serving vLLM On any Cloud with SkyPilot. Check out a 1-click example to start the vLLM demo, and the blog post for the story behind vLLM development on the clouds.
- [2023/06] We officially released vLLM! FastChat-vLLM integration has powered LMSYS Vicuna and Chatbot Arena since mid-April. Check out our blog post.
About
vLLM is a fast and easy-to-use library for LLM inference and serving.
vLLM is fast with:
- State-of-the-art serving throughput
- Efficient management of attention key and value memory with PagedAttention
- Continuous batching of incoming requests
- Fast model execution with CUDA/HIP graph
- Quantization: GPTQ, AWQ, SqueezeLLM, FP8 KV Cache
- Optimized CUDA kernels
vLLM is flexible and easy to use with:
- Seamless integration with popular Hugging Face models
- High-throughput serving with various decoding algorithms, including parallel sampling, beam search, and more
- Tensor parallelism support for distributed inference
- Streaming outputs
- OpenAI-compatible API server
- Support NVIDIA GPUs and AMD GPUs
- (Experimental) Prefix caching support
- (Experimental) Multi-lora support
vLLM seamlessly supports many Hugging Face models, including the following architectures:
- Aquila & Aquila2 (
BAAI/AquilaChat2-7B,BAAI/AquilaChat2-34B,BAAI/Aquila-7B,BAAI/AquilaChat-7B, etc.) - Baichuan & Baichuan2 (
baichuan-inc/Baichuan2-13B-Chat,baichuan-inc/Baichuan-7B, etc.) - BLOOM (
bigscience/bloom,bigscience/bloomz, etc.) - ChatGLM (
THUDM/chatglm2-6b,THUDM/chatglm3-6b, etc.) - Command-R (
CohereForAI/c4ai-command-r-v01, etc.) - DBRX (
databricks/dbrx-base,databricks/dbrx-instructetc.) - DeciLM (
Deci/DeciLM-7B,Deci/DeciLM-7B-instruct, etc.) - Falcon (
tiiuae/falcon-7b,tiiuae/falcon-40b,tiiuae/falcon-rw-7b, etc.) - Gemma (
google/gemma-2b,google/gemma-7b, etc.) - GPT-2 (
gpt2,gpt2-xl, etc.) - GPT BigCode (
bigcode/starcoder,bigcode/gpt_bigcode-santacoder, etc.) - GPT-J (
EleutherAI/gpt-j-6b,nomic-ai/gpt4all-j, etc.) - GPT-NeoX (
EleutherAI/gpt-neox-20b,databricks/dolly-v2-12b,stabilityai/stablelm-tuned-alpha-7b, etc.) - InternLM (
internlm/internlm-7b,internlm/internlm-chat-7b, etc.) - InternLM2 (
internlm/internlm2-7b,internlm/internlm2-chat-7b, etc.) - Jais (
core42/jais-13b,core42/jais-13b-chat,core42/jais-30b-v3,core42/jais-30b-chat-v3, etc.) - LLaMA, Llama 2, and Meta Llama 3 (
meta-llama/Meta-Llama-3-8B-Instruct,meta-llama/Meta-Llama-3-70B-Instruct,meta-llama/Llama-2-70b-hf,lmsys/vicuna-13b-v1.3,young-geng/koala,openlm-research/open_llama_13b, etc.) - MiniCPM (
openbmb/MiniCPM-2B-sft-bf16,openbmb/MiniCPM-2B-dpo-bf16, etc.) - Mistral (
mistralai/Mistral-7B-v0.1,mistralai/Mistral-7B-Instruct-v0.1, etc.) - Mixtral (
mistralai/Mixtral-8x7B-v0.1,mistralai/Mixtral-8x7B-Instruct-v0.1,mistral-community/Mixtral-8x22B-v0.1, etc.) - MPT (
mosaicml/mpt-7b,mosaicml/mpt-30b, etc.) - OLMo (
allenai/OLMo-1B,allenai/OLMo-7B, etc.) - OPT (
facebook/opt-66b,facebook/opt-iml-max-30b, etc.) - Orion (
OrionStarAI/Orion-14B-Base,OrionStarAI/Orion-14B-Chat, etc.) - Phi (
microsoft/phi-1_5,microsoft/phi-2, etc.) - Qwen (
Qwen/Qwen-7B,Qwen/Qwen-7B-Chat, etc.) - Qwen2 (
Qwen/Qwen1.5-7B,Qwen/Qwen1.5-7B-Chat, etc.) - Qwen2MoE (
Qwen/Qwen1.5-MoE-A2.7B,Qwen/Qwen1.5-MoE-A2.7B-Chat, etc.) - StableLM(
stabilityai/stablelm-3b-4e1t,stabilityai/stablelm-base-alpha-7b-v2, etc.) - Starcoder2(
bigcode/starcoder2-3b,bigcode/starcoder2-7b,bigcode/starcoder2-15b, etc.) - Xverse (
xverse/XVERSE-7B-Chat,xverse/XVERSE-13B-Chat,xverse/XVERSE-65B-Chat, etc.) - Yi (
01-ai/Yi-6B,01-ai/Yi-34B, etc.)
Install vLLM with pip or from source:
pip install vllm
Getting Started
Visit our documentation to get started.
Contributing
We welcome and value any contributions and collaborations. Please check out CONTRIBUTING.md for how to get involved.
Citation
If you use vLLM for your research, please cite our paper:
@inproceedings{kwon2023efficient,
title={Efficient Memory Management for Large Language Model Serving with PagedAttention},
author={Woosuk Kwon and Zhuohan Li and Siyuan Zhuang and Ying Sheng and Lianmin Zheng and Cody Hao Yu and Joseph E. Gonzalez and Hao Zhang and Ion Stoica},
booktitle={Proceedings of the ACM SIGOPS 29th Symposium on Operating Systems Principles},
year={2023}
}