Skip to content

Fix: Gemma4 checkpoint conversion failure when enable_nnx=True #4378

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

Merged
copybara-service merged 1 commit into from
Jul 8, 2026
Merged

Fix: Gemma4 checkpoint conversion failure when enable_nnx=True #4378

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
6 changes: 3 additions & 3 deletions src/maxtext/layers/nnx_decoders.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -653,7 +653,7 @@ def _init_scanned_gemma4(self, decoder_block_classes, rngs, mesh):
RemattedGemma4Block = gemma4.Gemma4ScannableBlock

if scan_length > 0:
self.layers = self._create_scanned_layers(
self.scanned_blocks = self._create_scanned_layers(
RemattedGemma4Block,
length=scan_length,
metadata_axis_name="layers",
Expand Down Expand Up @@ -2030,8 +2030,8 @@ def _apply_gemma4_scanned_blocks(
grouped_kv_caches = maxtext_utils.prepare_kv_caches_for_scan(
kv_caches, scan_length, attention_pattern_length, stack=False
)
y, self.layers, _ = self._apply_layers_sequentially(
self.layers, y, *layer_args, length=scan_length, kv_caches_stacked=grouped_kv_caches, **layer_kwargs
y, self.scanned_blocks, _ = self._apply_layers_sequentially(
self.scanned_blocks, y, *layer_args, length=scan_length, kv_caches_stacked=grouped_kv_caches, **layer_kwargs
)
maxtext_utils.update_kv_caches_after_scan(
kv_caches, grouped_kv_caches, scan_length, attention_pattern_length, stacked=False
Expand Down
Loading