Introduction
DeepSeek-R1-FlagOS-KunLunXin-INT8 provides an all-in-one deployment solution, enabling execution of DeepSeek-R1 on KunLunXin DSA. As the first-generation release for the KunLunXin-P800 series, this package delivers three key features:
- Comprehensive Integration:
- Integrated with FlagScale (https://github.com/FlagOpen/FlagScale).
- Open-source inference execution code, preconfigured with all necessary software and hardware settings.
- Pre-built Docker image for rapid deployment on KunLunXin-P800.
- INT8 Checkpoints:
- INT8 checkpoints generated by first dequantizing the official Deepseek-R1 FP8 checkpoint and then requantizing it to INT8.
- Consistency Validation:
- Evaluation tests verifying consistency of results between NVIDIA H100 and KunLunXin-P800.
Technical Summary
Serving Engine
We use FlagScale as the serving engine to improve the portability of distributed inference.
FlagScale is an end-to-end framework for large models across multiple chips, maximizing computational resource efficiency while ensuring model effectiveness. It ensures both ease of use and high performance for users when deploying models across different chip architectures:
- One-Click Service Deployment: FlagScale provides a unified and simple command execution mechanism, allowing users to fast deploy services seamlessly across various hardware platforms using the same command. This significantly reduces the entry barrier and enhances user experience.
- Automated Deployment Optimization: FlagScale automatically optimizes distributed parallel strategies based on the computational capabilities of different AI chips, ensuring optimal resource allocation and efficient utilization, thereby improving overall deployment performance.
- Automatic Operator Library Switching: Leveraging FlagScale's unified Runner mechanism and deep integration with FlagGems, users can seamlessly switch to the FlagGems operator library for inference by simply adding environment variables in the configuration file.
Triton Support
We validate the execution of DeepSeed-R1 model with a Triton-based operator library as a PyTorch alternative.
We use a variety of Triton-implemented operation kernels—approximately 60%—to run the DeepSeek-R1 model. These kernels come from two main sources:
Most Triton kernels are provided by FlagGems (https://github.com/FlagOpen/FlagGems), FlagGems is an operator library for large language models implemented in Triton. You can enable FlagGems kernels by setting the environment variable USE_FLAGGEMS. For more details, please refer to the "How to Run Locally" section.
Also included are Triton kernels from vLLM, including fused MoE.
INT8 Dequantization
We provide dequantized model weights in INT8 to run DeepSeek-R1 on KunLunXin GPUs, along with adapted configuration files and tokenizer.
bundle Download
Requested by Kunlunxin, the file of docker image and model files should be applied by email.
| Usage | Kunlunxin | |
|---|---|---|
| Basic Image | basic software environment that supports model running | [email protected] Contact by email，please indicate the unit/contact person/contact information/equipment source/specific requirements |
| Model | model weight and configuration files | [email protected] Contact by email，please indicate the unit/contact person/contact information/equipment source/specific requirements |
Contributing
We warmly welcome global developers to join us:
- Submit Issues to report problems
- Create Pull Requests to contribute code
- Improve technical documentation
- Expand hardware adaptation support
📞 Contact Us
Scan the QR code below to add our WeChat group send "FlagRelease"
License
This project and related model weights are licensed under the MIT License.