Yamz-Labs / kyojin
Kyojin: open-source engine that runs 100 GB-class open models (GLM-5.3, Qwen3.8, DeepSeek V4.1) on one desktop machine. AMD Strix Halo, NVIDIA RTX and Apple Mac. OpenAI API, built on ExLlamaV3.
Why this job: topics: apple-silicon, local-llm, llm-inference, inference, quantization
What it does
Strix Halo (AMD): install ROCm torch and build once (doc/install.md), then GLM-5.3-Flash:
Get started
git clone https://github.com/Yamz-Labs/kyojin && cd kyojin
hf download yamz-labs/GLM-5.3-Flash-EXL3-Yamz --local-dir ./glm-pack
python tools/glm/serve.py --model ./glm-pack --port 8000 -c 131072 --num-draft 2 - Maintenance
- Active
- Last commit today
- Stars
- 233
- +5 this week
- Licence
- MIT
- Allows business use
- Downloads for
- None listed
- No release files
- Latest version
- v1.8.14
- Released 2 days ago
- Built with
- Python
- 77% of the code
Is it looked after?
- Contributors
- 3
- Last 90 days
- Issues
- 90% closed
- 4 still open
- Pull requests
- 136 merged
- 2 waiting
- Releases
- 35
- 30+ in the last year
- Contributing guide (opens in new tab)
- No security policy
- Code of conduct (opens in new tab)
Similar repos
| Repo | Stars | Last commit | Licence | Downloads for |
|---|---|---|---|---|
| SemiAnalysisAI / InferenceX | 2,010 | today | Apache-2.0 | No downloads |
| LMCache / LMCache | 12,310 | today | Apache-2.0 | Linux |
| mw00 / project-maya | 2,712 | 2 days ago | MIT | No downloads |
More: languages, topics and project record
Languages
- Python 77%
- Cuda 16%
- C++ 4%
- HIP 1%
Topics
- amd
- exl3
- inference
- llm
- quantization
- rocm
- strix-halo
- apple-silicon
- cuda
- deepseek
- exllamav3
- glm
Project record
- JobLocal LLMs
- StatusListed
- Commits743
- Forks38
- Watching3
- Stars in 28 days+103
- CreatedOct 2, 2026
- First seen by usNov 7, 2026
- CheckedNov 9, 2026