Hugging Face's logo

Spaces:
sonic-coder
/
CPU-LLM-Inference
Sleeping

App Files Community
CPU-LLM-Inference

Commit History

  1. Update app.py

    Luigi
    Verified signature
    de64679

  2. Update app.py

    Luigi
    Verified signature
    4418827

  3. Update app.py

    Luigi
    Verified signature
    f2f4310

  4. Update app.py

    Luigi
    Verified signature
    42db70b

  5. feat(models): add Granite-4.0-Micro and Qwen3-4B-Instruct-2507 to MODELS registry⁠

    Luigi
    Verified signature
    c30a7f7

  6. feat(models): Added three new models⁠

    Luigi
    Verified signature
    3c22497

  1. add 5 models from liquid ai

    luigi-liu-tw
    8eefe94

  1. add smollm2 135m multilingual

    Luigi
    Verified signature
    ac20174

  1. add parser_model_ner_gemma_v0 based on gemma 3 370m it

    Luigi
    Verified signature
    bc1bd75

  1. Add Gemma-3-270m Taiwan

    Luigi
    Verified signature
    74c66e7

  1. add gemma-3-270m-it

    luigi-liu-tw
    3995aec

  1. remove prevously added breeze models (as it didn't work), add smollm 135m taiwan

    luigi-liu-tw
    b3fd72e

  1. add breeze models

    luigi-liu-tw
    88f3bc6

  1. add 3 sub-1B TW models from ShengweiPeng

    luigi-liu-tw
    ddfffab

  1. add Qwen2.5-Taiwan-3B-Reason-GRPO & Llama-3.2-Taiwan-1B

    luigi-liu-tw
    f82b9e0

  2. disable cache for better compability

    luigi-liu-tw
    ea7c44f

  1. make qwen2.5-taiwan 1.5b the default

    luigi-liu-tw
    493107e

  2. upgrade transforemrs to 4.53.3+

    luigi-liu-tw
    46df16c

  3. upgrade transformers to 4.46.1+

    luigi-liu-tw
    2a1d352

  4. add timm to requirements

    luigi-liu-tw
    0aa7a80

  5. add access_token to model loader

    luigi-liu-tw
    882aef2

  6. add gemma 3n, smallthinker 4b0.6b and falcon h1-1.5b

    luigi-liu-tw
    de303d7

  1. rename duckduckgo_search to ddgs

    luigi-liu-tw
    b858214

  2. remove qwen3 30ba3b & 32b, but add 14b

    luigi-liu-tw
    e883ef3

  3. use fp8 version of qwen3 30ba3b & 32b

    luigi-liu-tw
    9ece009

  4. add autoawq

    luigi-liu-tw
    df9e599

  5. add qwe3 30ba3b awq & 30b awq

    luigi-liu-tw
    3a690c1

  1. update prompt

    luigi-liu-tw
    8d42201

  2. update prompt

    luigi-liu-tw
    5398786

  3. fix err in prompt

    luigi-liu-tw
    b1fb4c7

  4. improve prompt for better web search result merging

    luigi-liu-tw
    6a87eb4

  5. Add D-R1-0528 8B, Nemotron 1.5B, Qwen2.5 Taiwan 1.5B

    luigi-liu-tw
    12d3fe3

  1. reduce rep penality to 1.2

    luigi-liu-tw
    079e166

  2. increase rep penality as on many slms, reptition is frequent.

    luigi-liu-tw
    8a7c97d

  3. revert device_map to audo

    luigi-liu-tw
    cc12a51

  4. Add smolLM2-135M GRPO and SmolLM2-360M

    luigi-liu-tw
    cb16ac6

  1. Update README.md

    Luigi
    Verified signature
    371669a

  2. update readme

    luigi-liu-tw
    076c1f2

  1. add Qwen2.5-Omni-3B & MiMo-7B-RL

    luigi-liu-tw
    6a4537b

  2. disable models unrunnable on hf spaces

    luigi-liu-tw
    7c5f318

  3. switch Phi-4-mini-Instruct" from unsloth to microsoft

    luigi-liu-tw
    23b3848

  4. extend max tokens

    luigi-liu-tw
    d730ffe

  5. adjust layout

    luigi-liu-tw
    4911925

  6. add "Granite-4.0-Tiny-Preview" model

    luigi-liu-tw
    09d9700

  7. add phi-4 reasonning and phi-4-mini reasonning

    luigi-liu-tw
    68e6569

  8. bugfix: set device to xpu by mistake

    luigi-liu-tw
    1242438

  9. add all qwen3 variants

    luigi-liu-tw
    2882063

  10. user can define search timeout

    luigi-liu-tw
    e2ee907

  11. give 5 second for web earch to gather reults

    luigi-liu-tw
    c00d442

  12. support thinking models and streamingly display thought

    luigi-liu-tw
    8c3c2b9