DFlash Qwen3-0.6B — 1024 len, batch 12, 6 epochs
First full 1024-length DFlash run on the fast-iteration setup. Reached val/greedy_accepted_length 4.770; validation was still rising at the end.
Workflow
Experiment description success finished 59d 0h ago #
What this experiment is: a baseline of its own, or a diff on top of an existing run. Carries the repository, the baseline it is measured against, and the setup it reuses.
GPU server success finished 59d 0h ago #
Which machine or cluster runs the training, and what hardware it provides. Chosen manually for now.
No artifacts recorded yet.
Experiment preparation success #
Get code, environment, and data onto the chosen server.
Copy code to server success finished 59d 0h ago #
No artifacts recorded yet.
Create venv and install prerequisites success finished 59d 0h ago #
No artifacts recorded yet.
Download data and preprocess data cache success finished 59d 0h ago #
No artifacts recorded yet.
Other preparation success finished 59d 0h ago #
No artifacts recorded yet.
Sanity checks success #
Cheap checks that must pass before spending GPU hours on a full run.
Unit and integration tests success finished 59d 0h ago #
Cover the newly added code. For a diff on top of a baseline, run the affected tests rather than the full suite.
No artifacts recorded yet.
Smoke test success finished 59d 0h ago #
A handful of training steps on a small setup with debug hyperparameters. Must finish within about 7 minutes.
No artifacts recorded yet.
Full training success #
The real training run and its health over time.
Launch and watch first steps success finished 59d 0h ago #
Start the full run and watch the first training steps to confirm it behaves like the smoke test did.
No artifacts recorded yet.
Scheduled monitoring success finished 59d 0h ago #
Every 30 minutes, check the run is alive and its metrics are healthy. Restart it if either check fails.
Full evaluation success #
The evaluation run, monitored the same way as training.
Launch and watch first steps success finished 59d 0h ago #
Start the evaluation and watch the first steps to confirm it is producing sane numbers.
No artifacts recorded yet.
Scheduled monitoring success finished 59d 0h ago #
Every 30 minutes, check the evaluation is alive and healthy. Restart it if either check fails.