Commit History

  1. Trigger rebuild for demospace

    b5fcd1b
  2. Update README

    03dbec2
  3. Update space

    0b786ba
  4. Update backend with Google OAuth and refresh documentation

    d1a2346
  5. Update space

    d680311

  1. Added security checks, updated README.md to reflect that

    0986c85

  1. Finalized changes, minimal clean-up

    733733f
  2. Fixed chat permanency

    053c6ef
  3. UI FIX

    1c37cc9
  4. UI FIX

    ec837f4
  5. fixed prompt build again

    c1df21f
  6. fixed prompt build

    a29613f
  7. changed websearch api from duckduckgo to tavily AGAIN

    933dfec
  8. changed websearch api from duckduckgo to tavily

    30c33af
  9. fixed requirements

    4718379
  10. downgraded model to fit ram constrains

    9cc58a7
  11. fixed web search

    f802a2e
  12. changed parse user mode, making it easier for the code to detect casual talk

    1ebdff9

  1. resource tunning

    a05885e
  2. tunned prompt

    d51c437
  3. turned off f16_kv

    04eb1b4
  4. Reduced max tokens from 1024 to 512

    da3f013
  5. Reduced n_ctx from 8000 to 4000

    d1b5104
  6. Fixed web search and file upload

    27a4db8
  7. Changed README to reflect code changes

    37fad88
  8. Finetuning model and parameters again, ram increase

    5f428fc
  9. Finetuning model and parameters

    781cd42
  10. Finetuning model and parameters

    6999581
  11. fix: returning to full production app.py

    3fc17e3
  12. fix: testing stream

    ed56a60
  13. fix: added stay alive to stream

    4d9458a
  14. fix: readloop

    196b9ca
  15. fixed generate function

    a18e17a
  16. fixed f-string error

    2c3628a
  17. fixed minimal stream route

    5a745cf
  18. changed to a smaller qwen model

    0e8eb8a
  19. fix: align prompt budget with context window and add keep-alive to stream

    03a5dae
  20. fix: resolve NameError for MODEL and optimize RAM context

    c06aafd

  1. fix: resolve NameError for MODEL and optimize RAM context

    389d7ea
  2. fix: remove ollama dependency and use self-contained llama-cpp

    19ecab1
  3. fix: update gunicorn path to find app in backend folder

    433d501
  4. Initial clean production-ready commit

    f17b878