C++ · MIT
sergqwer/strata-nvfp4
Strata fork: Qwen3.8-Flash-Next 125B MoE in NVFP4 on one RTX 20-50 card (12 GB+) and 64 GB of RAM or more. Setup installs our GPTQ quants (huihui abliterated, OrcaRouter) from Hugging Face. CPU experts on AVX-512 or AVX2 (+AVX-VNNI), all experts in RAM from 92 GB, a low-RAM mode below, vision, W4A8 prompts. Windows / CUDA 13.
★ 221
Stars
26 forks · 16 open issues
+3.0/day
Growth · measured
Stars counted between two of our own snapshots roughly a week apart. The repository page gives the exact span.
32.3
Momentum / 100
see the breakdown below
39d ago
Created
last push 14d ago
Star history
How this score was built
| Component | Weight | Earned | Why |
|---|---|---|---|
| Velocity | 55 | 11.5 | +3.0/day, log-scaled against a 750/day ceiling |
| Acceleration | 15 | 0.0 | last week ran 0.68× the monthly pace |
| Freshness | 15 | 14.0 | last push 14d ago; decays to zero at 180 days |
| Scale | 15 | 6.8 | 221 stars, log-scaled against 150k — stops a 3-star repo topping the board |
| Momentum | 100 | 32.3 | sum of the above |
Detail
| Topics | avx2avx512blackwellcudagptqllm-inferencemoenvfp4qwenvisionwindows |
| Matched on | LLM & Inference (llm-inference) |
| Short window | +21 stars over 7 days = 3.0/day |
| Long window | +114 stars over 26 days = 4.4/day |
| Lifetime average | 6.00 stars/day over 37 days |
| First seen here | 2026-10-11 |
| Last read | 2d ago |