guoqingbao / xinfer
Blazing-fast LLM inference in pure Rust. No PyTorch and Python runtime.
Why this job: says: llm inference; topics: vllm
What it does
MoE expert offloading is supported on CUDA only. Run a large MoE model with expert weights offloaded to host RAM. The value is a fraction: 0.2 keeps 20% of each layer's experts on the GPU; auto selects the cache size automatically. Tensor parallelism requires NCCL.
Get started
curl -sSL https://guoqingbao.github.io/xinfer/install.sh | bash - Maintenance
- Active
- Last commit 10 days ago
- Stars
- 347
- +2 this week
- Licence
- MIT
- Allows business use
- Downloads for
- 2 systems
- Mac, Linux
- Latest version
- v0.14.7
- Released 9 days ago
- Built with
- Rust
- 88% of the code
Is it looked after?
- Contributors
- Not stated
- Last 90 days
- Issues
- 81% closed
- 48 still open
- Pull requests
- 202 merged
- 15 waiting
- Releases
- 37
- 30+ in the last year
- No contributing guide
- No security policy
- No code of conduct
Similar repos
| Repo | Stars | Last commit | Licence | Downloads for |
|---|---|---|---|---|
| pegainfer-project / pegainfer | 738 | yesterday | Apache-2.0 | Linux |
| xorbitsai / inference | 9,647 | yesterday | Apache-2.0 | No downloads |
| microsoft / vidur | 721 | 2 months ago | MIT | No downloads |
More: download files, languages, topics and project record
Download files, v0.14.7
- Mac2 files, 5 downloads
- Linux24 files, 69 downloads
Languages
- Rust 88%
- HTML 6%
- Python 3%
- CSS 1%
Topics
- agent
- llm
- qwen
- rust
- vllm
Project record
- JobLocal LLMs
- StatusListed
- Commits493
- Forks52
- Watching5
- Stars in 28 days+14
- CreatedJun 19, 2025
- First seen by usNov 7, 2026
- CheckedNov 8, 2026