Commit History

  1. POCKET family cross-links + Image/Studio/Zimage

    Verified signature
    0b3b76f
  2. Link POCKET family + cross-references

    Verified signature
    a0bac4b

  1. A/B: raise max_tokens 256->512 so POCKET's reasoning + final answer complete (not cut)

    Verified signature
    336526e
  2. A/B: show reasoning again (remove no_think prefill); raise A/B max_tokens 96->256 so thinking completes

    Verified signature
    8bfa67a
  3. A/B: one-sentence example prompts (stable tok/s, still fast)

    Verified signature
    7bf27a9
  4. A/B: prefill empty <think></think> to force direct short answers (no reasoning) -> fast races

    Verified signature
    d6e52f6
  5. A/B: apply /no_think so both models answer directly (no <think> block) -> short, fast races

    Verified signature
    f8feff1
  6. A/B: shorter-answer example prompts + cap max_tokens 200->96 for snappy races

    Verified signature
    87a05d4
  7. A/B is now the landing tab (swap order); add prism-ml/Bonsai-27B-gguf to models relation

    Verified signature
    bb234a8
  8. add Tab2 live A/B vs Bonsai (sequential: Bonsai first, then POCKET); 2nd llama-server for Bonsai Q1_0; README=BONSAI vs POCKET

    Verified signature
    e77d0bf
  9. show live tok/s: backend streams predicted_per_second; UI displays live speed + final tok/s under each reply

    Verified signature
    4df8486
  10. proxy via native /completion with manual ChatML prompt (bypass embedded template); surface backend errors

    Verified signature
    ea6a70b
  11. use prebuilt llama.cpp release b10091 (qwen35moe, runtime CPU dispatch) - no compile (no OOM), no -march=native (no SIGILL)

    Verified signature
    6444aa4
  12. fix: cap compile parallelism (-j2) to avoid OOMKilled (exit137) during llama.cpp build

    Verified signature
    87e88c2
  13. rebuild backend on upstream llama.cpp master (qwen35moe support) via llama-server; FastAPI proxies; portable CPU build

    Verified signature
    b07a776
  14. fix: portable CPU build (no -march=native/AVX512) -> stop SIGILL exit132; graceful load-error status

    Verified signature
    2ce7d5d
  15. fix: portable CPU build (no -march=native/AVX512) -> stop SIGILL exit132; graceful load-error status

    Verified signature
    5184a79
  16. Upload README.md with huggingface_hub

    Verified signature
    be67c4f
  17. Upload README.md with huggingface_hub

    Verified signature
    40d655e
  18. Upload README.md with huggingface_hub

    Verified signature
    9b4c19e
  19. Upload requirements.txt with huggingface_hub

    Verified signature
    11db75e
  20. Upload Dockerfile with huggingface_hub

    Verified signature
    f4eae4e
  21. Upload index.html with huggingface_hub

    Verified signature
    1e1f8f3
  22. Upload app.py with huggingface_hub

    Verified signature
    e186935
  23. initial commit

    Verified signature
    fef847f