Deepseek R1 AWQ

🚀 DeepSeek R1 AWQ

This project offers the AWQ quantization of DeepSeek R1, aiming to optimize the model's performance and resource utilization.

🚀 Quick Start

To serve using vLLM with 8x 80GB GPUs, use the following command:

VLLM_USE_V1=0 VLLM_WORKER_MULTIPROC_METHOD=spawn VLLM_MARLIN_USE_ATOMIC_ADD=1 python -m vllm.entrypoints.openai.api_server --host 0.0.0.0 --port 12345 --max-model-len 65536 --max-seq-len-to-capture 65536 --enable-chunked-prefill --enable-prefix-caching --trust-remote-code --tensor-parallel-size 8 --gpu-memory-utilization 0.95 --served-model-name deepseek-reasoner --model cognitivecomputations/DeepSeek-R1-AWQ

You can download the wheel I built for PyTorch 2.6, Python 3.12 by clicking here. The benchmark below was done with this wheel, it contains 2 PR merges and an unoptimized FlashMLA (still faster than Triton) for A100 which boosted performance a lot. The vLLM repo which contained A100 FlashMLA can be found at LagPixelLOL/vllm@sm80_flashmla, which is a fork of vllm-project/vllm. The A100 FlashMLA it used is based on LagPixelLOL/FlashMLA@vllm, which is a fork of pzhao-eng/FlashMLA.

✨ Features

• Quantization: Quantized by Eric Hartford and v2ray. This quant modified some of the model code to fix an overflow issue when using float16.
• Performance Boost: The unoptimized FlashMLA for A100 in the provided wheel significantly boosts performance.

📚 Documentation

TPS Per Request

Notes

• A100 Config: The A100 config uses an unoptimized FlashMLA implementation, which is only superior than Triton during high context inference. It would be faster if it's optimized.
• L40S Config: The L40S config doesn't support FlashMLA, thus the Triton implementation is used. This makes it extremely slow with high context. But the L40S doesn't have much VRAM, so it can't really have that much context anyway, and it also doesn't have the fast GPU to GPU interconnection bandwidth, making it even slower. It is not recommended to serve with this config, as you must limit the context to <= 4096, --gpu-memory-utilization to 0.98, and --max-num-seqs to 4.
• GPU Form Factor: All types of GPU used during benchmark are SXM form factor except L40S.
• Inference Speed: Inference speed will be better than FP8 at low batch size but worse than FP8 at high batch size, this is the nature of low bit quantization.
• vLLM Support: vLLM supports MLA for AWQ now, you can run this model with full context length on just 8x 80GB GPUs.

📄 License

This project is licensed under the MIT license.