Python · MIT
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.
★ 229
Stars
38 forks · 2 open issues
+1.6/day
Growth · measured
Stars counted between two of our own snapshots roughly a week apart. The repository page gives the exact span.
29.7
Momentum / 100
see the breakdown below
36d ago
Created
last push 2d ago
Star history
How this score was built
| Component | Weight | Earned | Why |
|---|---|---|---|
| Velocity | 55 | 7.9 | +1.6/day, log-scaled against a 750/day ceiling |
| Acceleration | 15 | 0.0 | last week ran 0.39× the monthly pace |
| Freshness | 15 | 15.0 | last push 2d ago; decays to zero at 180 days |
| Scale | 15 | 6.8 | 229 stars, log-scaled against 150k — stops a 3-star repo topping the board |
| Momentum | 100 | 29.7 | sum of the above |
Detail
| Topics | amdapple-siliconcudadeepseekexl3exllamav3glminferencellmllm-inferencelocal-llmnvidiaopenai-apiquantizationqwenrocmstrix-halo |
| Matched on | LLM & Inference (llm, llm-inference) |
| Short window | +11 stars over 7 days = 1.6/day |
| Long window | +92 stars over 23 days = 4.0/day |
| Lifetime average | 6.81 stars/day over 34 days |
| First seen here | 2026-10-13 |
| Last read | 2d ago |