ji

docs/AUDIT.md

@@ -494,9 +494,12 @@ The overnight scheduler runs `run_pipeline()` via `asyncio.to_thread`; the threa
 on the training subprocess for hours. Cancelling the scheduler task (quit) released
 the training lock immediately while the thread kept training, and at interpreter exit
 Python joins executor threads — so Javis hung on quit until training finished.
-**Fix:** cancellation calls `piper_train.request_stop()`, which terminates the child
-process (`_run` now polls instead of blocking in `wait()`); the lock is only released
-once the worker has actually finished.
+**Fix:** cancellation calls `piper_train.request_stop()`: the training process finishes
+its step, saves its progress — within the epoch too: the next run continues with the
+next batch — and exits (it is killed if that takes longer than
+`STOP_GRACE_SECONDS`), and the scheduler waits for that (`_run` polls instead of
+blocking in `wait()`). The lock (`training/.training.lock`, held by `run_pipeline`)
+names the training processes and is only released once they have exited.
 
 #### VOI-01
 **Imported voice archives may carry pickle checkpoints** — *Medium (trust)*