feat: use Gemini response metadata for token counting #11743
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
dosubot
bot
added
size:XS
|
Please fix the lint. |
|
@crazywoola |
36 tasks
crazywoola
approved these changes
Dec 17, 2024
tastelikefeet
added a commit
to tastelikefeet/dify
that referenced
this pull request
Dec 19, 2024
…elscope * commit '2624a6dcd0b89dd7c1aac2d7bfe7f769e9e3c992': (630 commits) Fix explore app icon (langgenius#11808) ci: fix config ci and it works (langgenius#11807) ci: add config ci more disscuss check langgenius#11706 (langgenius#11752) chore: bump version to 0.14.1 (langgenius#11784) feat:add hunyuan model(hunyuan-role, hunyuan-large, hunyuan-large-rol… (langgenius#11766) chore(opendal_storage): remove unused comment (langgenius#11783) feat: Disable the "Forgot your password?" button when the mail server setup is incomplete (langgenius#11653) chore(.env.example): add comments for opendal (langgenius#11778) Lindorm vdb bug-fix (langgenius#11790) fix: imperfect service-api introduction text (langgenius#11782) feat: add openai o1 & update pricing and max_token of other models (langgenius#11780) fix: file upload auth (langgenius#11774) feat: add parameters for JinaReaderTool (langgenius#11613) feat: full support for opendal and sync configurations between .env and docker-compose (langgenius#11754) feat(app_factory): speed up api startup (langgenius#11762) fix: Prevent redirection to /overview when accessing /workflow. (langgenius#11733) (doc) fix: update cURL examples to include Authorization header (langgenius#11750) Fix explore app icon (langgenius#11742) chore: improve gemini models (langgenius#11745) feat: use Gemini response metadata for token counting (langgenius#11743) ...
Scorpion1221
added a commit
to yybht155/dify
that referenced
this pull request
Dec 19, 2024
* commit '926546b153a701dfbfc71eb8157f9a41320444f8': (486 commits) chore: bump version to 0.14.1 (langgenius#11784) feat:add hunyuan model(hunyuan-role, hunyuan-large, hunyuan-large-rol… (langgenius#11766) chore(opendal_storage): remove unused comment (langgenius#11783) feat: Disable the "Forgot your password?" button when the mail server setup is incomplete (langgenius#11653) chore(.env.example): add comments for opendal (langgenius#11778) Lindorm vdb bug-fix (langgenius#11790) fix: imperfect service-api introduction text (langgenius#11782) feat: add openai o1 & update pricing and max_token of other models (langgenius#11780) fix: file upload auth (langgenius#11774) feat: add parameters for JinaReaderTool (langgenius#11613) feat: full support for opendal and sync configurations between .env and docker-compose (langgenius#11754) feat(app_factory): speed up api startup (langgenius#11762) fix: Prevent redirection to /overview when accessing /workflow. (langgenius#11733) (doc) fix: update cURL examples to include Authorization header (langgenius#11750) Fix explore app icon (langgenius#11742) chore: improve gemini models (langgenius#11745) feat: use Gemini response metadata for token counting (langgenius#11743) chore: update comments in docker env file (langgenius#11705) feat(ark): support doubao vision series models (langgenius#11740) chore: the consistency of MultiModalPromptMessageContent (langgenius#11721) ... # Conflicts: # api/configs/app_config.py # api/core/helper/code_executor/code_executor.py # web/yarn.lock
jiangbo721
pushed a commit
to jiangbo721/dify
that referenced
this pull request
Dec 20, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
Previously, I fixed it in another PR(#11226), but the streaming function wasn't corrected, so this is an additional fix.
Improved token counting for Gemini models by utilizing response metadata. Previously, the system used GPT-2 tokenizer which resulted in inaccurate token counts for both prompt and completion. Now, the system first attempts to get token counts from Gemini's response metadata, falling back to manual calculation only when metadata is unavailable. This change ensures accurate token counting and improves efficiency by leveraging native Gemini functionality.
Tip
Close issue syntax:
Fixes #<issue number>orResolves #<issue number>, see documentation for more details.Screenshots
Checklist
Important
Please review the checklist below before submitting your pull request.
dev/reformat(backend) andcd web && npx lint-staged(frontend) to appease the lint gods