Code

History

Commits on example-docs to crates/edit_prediction_cli35+
  1. ep: Ensure prompt is always within token limit (#51529)Ben Kunkle committed 8 months ago8a7daea
  2. ep: Add multi-region format (#51185)Oleksiy Syvokon committed 8 months ago376d410
  3. ep: Include diagnostics in `ZetaPromptInput` (#51141)Ben Kunkle committed 8 months ago1475774
  4. zeta: Allow the server to select the editable and context ranges more flexibly (#50975)Max Brunsfeld committed 8 months ago2bd5c21
  5. ep: Skip context retrieval when already performed (#51100)Oleksiy Syvokon committed 8 months agob54716d
  6. ep: Make repair parser more robust (#51044)Oleksiy Syvokon committed 8 months agoaf2b35a
  7. ep: Add captured example fetching (#50960)Ben Kunkle committed 8 months ago5b2d39e
  8. ep: Unify code in `parse_output` and `zeta_prompt` (#50958)Oleksiy Syvokon committed 8 months ago4b3660b
  9. Refresh LLM API token on organization change (#50931)Neel committed 8 months agodc0e41f
  10. zeta: Add variable edit format (#50850)Max Brunsfeld committed 8 months ago61d9696
  11. zeta2: Don't remove redundant excerpts on the client (#50886)Ben Kunkle committed 8 months agoda9d548
  12. ep: Make --repetitions work with Anthropic and OpenAI batches (#50789)Oleksiy Syvokon committed 8 months ago2991b05
  13. Include repo_url in edit prediction requests when data collection is enabled (#50745)Max Brunsfeld committed 8 months ago67c2c34
  14. ep: Collapse whitespace in deltaChrF (#50716)Oleksiy Syvokon committed 8 months ago0394341
  15. ep: Option to configure custom Baseten environment (#50706)Oleksiy Syvokon committed 8 months ago9b8ad01
  16. zeta2: Hashlines prompt format (#50623)Ben Kunkle committed 8 months agodc7cbd3
  17. ep: Predict by querying Baseten directly (#50626)Oleksiy Syvokon committed 8 months ago09e178d
  18. ep: Fix fetching rated-after (#50617)Oleksiy Syvokon committed 8 months ago528bf7c
  19. Remove Zeta1 logic, allow choosing zeta2 experiment (#50560)Max Brunsfeld committed 8 months ago880b2e5
  20. ep_cli: Clean up snowflake fetching (#50450)Ben Kunkle committed 8 months ago7eb7458
  21. ep: Add settled data fetching from snowflake (#50326)Ben Kunkle committed 8 months agoceb9d83
  22. Add edit prediction evals that test related excerpt usage (#50256)Max Brunsfeld committed 8 months agod4e89f9
  23. Add fast mode toggle in agent panel (#49714)Tom Houlé committed 8 months ago6a74938
  24. Edit prediction: add some typescript evals, fix a teacher model failure mode (#50141)Max Brunsfeld committed 8 months ago3da9adc
  25. ep: Stratify by cursor_path by default (#50111)Oleksiy Syvokon committed 8 months agoc9aea6f
  26. Fix incorrect cursor excerpt handling in EP CLI format-prompt (#49947)Max Brunsfeld committed 8 months ago096610e
  27. Fix handling of excerpt regions in EP CLI (#49936)Max Brunsfeld committed 8 months ago5ef898d
  28. zeta: Consolidate logic for picking region sizes, use larger editable region (#49921)Max Brunsfeld committed 8 months ago97a8fe7
  29. ep: Add a parameter to sample at most N near-duplicates (#49870)Oleksiy Syvokon committed 8 months ago450c66c
  30. ep: Fix teacher parser and other fixes (#49863)Oleksiy Syvokon committed 8 months ago67d9e7d
  31. ep_cli: Only compute reversals on last user edit (#49747)Ben Kunkle committed 8 months ago8f5ae2d
  32. ep_cli: Delete sync-deployments command (#49728)Ben Kunkle committed 8 months agod05ea29
  33. ep_cli: Don't filter rated predictions by Zed version (#49609)Ben Kunkle committed 8 months agod896bf7
  34. Allow using Zeta through an arbitrary OpenAI-compatible self-hosted API (#49554)Max Brunsfeld committed 8 months ago62af5b8
  35. gpui: Extract gpui_platform out of gpui (#49277)Piotr Osiewicz committed 8 months agobc31ad4