Hugging Face's logo

Spaces:
Leon4gr45
/
fable5-inference
Running

App Files Community
fable5-inference

Commit History

Yesterday

  1. Upload folder using huggingface_hub Running

    Leon4gr45
    Verified signature
    764508f

  2. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    7bdae9c

  3. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    d56d887

  4. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    cf352b2

  5. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    db537b1

  6. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    c141489

  7. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    e4a5d22

  8. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    93c9ace

  9. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    d15175a

  10. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    449992f

  11. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    674c3c5

  12. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    0eda7e2

  13. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    8189e73

  14. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    74a58b1

  15. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    24a31e9

  16. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    b6caecd

  17. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    a309d45

  18. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    78aed6c

  19. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    bee04df

  20. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    d1c689e

  21. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    62cb678

2 days ago

  1. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    209ba69

  2. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    93e8674

  3. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    02abefa

  4. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    2acc0a6

  5. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    0d74b42

3 days ago

  1. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    bce7cab

  2. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    d8f8e80

  3. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    b92eac6

  4. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    b4ef077

  5. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    df1ad21

  6. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    a389da3

5 days ago

  1. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    6aa9bb1

6 days ago

  1. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    46a7771

  1. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    fd44567

  2. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    de4c99c

  3. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    5eb4bb2

  4. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    b8ef05e

  5. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    c7271de

  1. Document split KV quant + cache-reuse env vars

    Leon4gr45
    Verified signature
    4963346

  2. Split K/V cache quant (K=q8_0,V=q4_0), 32K ctx, cache-reuse, tool-call smoke test

    Leon4gr45
    Verified signature
    b6dd083

  1. safety: cap threads at CPU_THREADS_MAX (default 2) to prevent oversubscription if detection over-reports

    Leon4gr45
    Verified signature
    111e96a

  2. feat: cgroup-aware vCPU auto-detection (respects container CPU limit; no CPU_THREADS env needed)

    Leon4gr45
    Verified signature
    1256f93

  3. fix: startup readiness loop must sleep on HTTP 503 (model loading), not only on connection errors — false 'startup timeout' left instance marked unloaded

    Leon4gr45
    Verified signature
    261a815

  4. harden: normalize legacy KV_CACHE_QUANT values (4bit->q4_0) so stale env vars can't crash llama-server

    Leon4gr45
    Verified signature
    ea909c9

  5. build: use prebuilt llama.cpp CPU binary (b9895) instead of source compile — fixes mtmd OOM hang, ~20x faster build

    Leon4gr45
    Verified signature
    ea5d56a

  6. build: pull model at container runtime instead of baking into image (faster/reliable build)

    Leon4gr45
    Verified signature
    7cc60cb

  7. v2.0: single-instance CPU proxy, native OpenAI tools+streaming, flash-attn+q4_0 KV, all-cores

    Leon4gr45
    Verified signature
    e24791f

  8. Fix llama-server arguments and increase timeouts

    Leon4gr45
    Verified signature
    1bc53ac

  9. Upload folder using huggingface_hub

    Leon4gr45
    Verified signature
    912716f