Faster quiz generation: planned parallel batches, room to reason
All checks were successful
Verify and publish V2 container / build-and-push (push) Successful in 25s
All checks were successful
Verify and publish V2 container / build-and-push (push) Successful in 25s
- Output limits leave room for reasoning (answer plus 8k-64k by effort, up to the model maximum); a call cut off before answering is retried with double the limit. At "max" effort, batches were being cut off at 16k tokens. - Quizzes are planned first (one distinct angle per question), then written in waves of four batches; med math runs alongside the first wave. - Replacements for removed questions are planned and written in parallel. - The Python check and the second review run at the same time each round. Measured live (DeepSeek flash, max reasoning, 50 questions): 7 min 21 s with no duplicates, against 20 min and 29 replaced duplicates without planning. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
parent
1a212bb8ba
commit
2454ab6e95
8 changed files with 213 additions and 51 deletions
12
docs/AI.md
12
docs/AI.md
|
|
@ -105,6 +105,18 @@ has no calculations. Topics are kept broad ("Stroke", not "Right hemisphere stro
|
|||
prompt lists the class's existing topics, and narrower labels that contain an existing one are
|
||||
merged into it after generation.
|
||||
|
||||
Speed: a quiz of more than one batch is planned first. One call lists a distinct angle per
|
||||
question (client situation plus the single decision or fact tested), then the batches of 8
|
||||
run four at a time (`WRITE_WAVE`), each writing its own angles from its own excerpts, so
|
||||
parallel batches do not repeat each other. Requested med math is written alongside the first
|
||||
wave. Replacements for removed questions are planned and written the same way. Output limits
|
||||
leave room for reasoning (`outputBudget` in `server/ai/service.ts`: the answer plus 8k-64k by
|
||||
reasoning effort, up to the model's maximum, unless a limit is set in Settings), and a call cut
|
||||
off before answering is retried with double the limit. In each check round the Python check and
|
||||
the second review run at the same time. Measured on DeepSeek flash at "max" reasoning, a
|
||||
50-question quiz from a one-paragraph exam took 7 min 21 s (no duplicates, two rounds); before,
|
||||
calls ran out of output space and the same quiz did not finish in 15 minutes.
|
||||
|
||||
Flashcards are atomic: one fact per card, the back only the answer (a word, number or short
|
||||
phrase), and no hints (the schema has no hint field any more). When generation finishes, one
|
||||
small call names the set: a 2-6 word title for what it covers, different from the class's other
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue