DeepInfer
DeepInfer: High-performance, large-scale inference framework for DeepSeek models
DeepInfer launched on 16 October 2026. It ranked #42 on Daily Launch on 30 October 2026 with a Launch Score of 45, from +293 stars today on GitHub. It moved down 14 places from the day before.
What is DeepInfer?
DeepInfer is the high-performance inference framework that powers DeepSeek's large-scale production services. This open-source release provides offline inference with the core features needed for serving at scale. Before building DeepInfer, please create and activate an empty Python 3.12 environment. The default checkpoint, deepseek-ai/DeepSeek-V4.1-Flash-1017, is downloaded into the Hugging Face cache on the first run; set DEEPINFERHFMODELDIR=/path/to/DeepSeek-V4.1-Flash-1017 to run from a local copy instead.
From its README on GitHub
DeepInfer is a repo by deepseek-ai on GitHub. It is written in Python and released under the Apache-2.0 license.
How to get DeepInfer
Get it from its repository on GitHub.