Skip to content

Fix checkpoint saving using wrong interval config #320

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Mr-Neutr0n wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Fix checkpoint saving using wrong interval config #320

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2826b63
Fix checkpoint saving using wrong interval config
Mr-Neutr0n Feb 11, 2026
4eee69e
test: cover checkpoint interval scheduling
Mr-Neutr0n Jul 31, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 7 additions & 2 deletions vit_jax/train.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,6 +34,12 @@
from vit_jax import utils


def _should_checkpoint(config, step, total_steps):
"""Returns whether a checkpoint should be stored at the current step."""
return (step == total_steps or
bool(config.checkpoint_every and step % config.checkpoint_every == 0))


def make_update_fn(*, apply_fn, accum_steps, tx):
"""Returns update step for data parallel training."""

Expand Down Expand Up @@ -239,8 +245,7 @@ def init_model():
img_sec_core_test=img_sec_core_test))

# Store checkpoint.
if ((config.checkpoint_every and step % config.eval_every == 0) or
step == total_steps):
if _should_checkpoint(config, step, total_steps):
checkpoint_path = flax_checkpoints.save_checkpoint(
workdir, (flax.jax_utils.unreplicate(params_repl),
flax.jax_utils.unreplicate(opt_state_repl), step), step)
Expand Down
15 changes: 15 additions & 0 deletions vit_jax/train_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,6 +36,21 @@

class TrainTest(parameterized.TestCase):

@parameterized.named_parameters(
('configured_interval', 6, 10, 3, True),
('between_intervals', 4, 10, 3, False),
('disabled', 4, 10, 0, False),
('final_step_when_disabled', 10, 10, 0, True),
)
def test_should_checkpoint(self, step, total_steps, checkpoint_every,
expected):
config = ml_collections.ConfigDict({
'checkpoint_every': checkpoint_every,
'eval_every': 2,
})
self.assertEqual(
train._should_checkpoint(config, step, total_steps), expected)

@parameterized.named_parameters(
('tfds', 'tfds'),
('directory', 'directory'),
Expand Down