Repository navigation
Conversation
- Ask for a "statement" field and a True/False statement package, so small models stop returning questions; reject items that are still phrased as questions. - When a generated package is rejected, regenerate it once and tell the model what was wrong, and log the reason as practice_question_validation_failed. - Credit evidence quoted verbatim from a different supplied passage to that passage, and ignore stray PDF spaces before punctuation.
- Ask Ollama to keep the model loaded for 10 minutes after each request. - Load the model in the background when Chat or Practice opens, and refresh it on activity at most every 4 minutes, so returning after a pause starts loading before the next request. Refreshing reuses the loaded context size so it never triggers a reload. - Unload the models the app used when it quits.
- Record each checked True/False answer under the session's topic and derive mastery from the attempt log: a correct answer counts 1 (0.5 with help), an incorrect answer counts 1 against. - Show mastery in the topic bar, the progress review, and a "Topics to revisit" list on the topic picker, weakest first. - Schema 13: one practice_attempts table. Databases with the schema 12 attempt log and topic_mastery table are migrated.
questions count more towards mastery than older questions.
This file contains hidden or 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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Added 3 commits.
The first ensures that the true/false question generation actually generates true/false questions instead of generating normal questions.
The second ensures that the model stays loaded between chat and practice mode.
The third adds mastery tracking to the new practice mode.