Skip to content

[CI] Fix Nemotron nightly mixed precision checkpoints test #27284

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
Fridge003 merged 2 commits into from
Jun 5, 2026
+8 −6
Merged

[CI] Fix Nemotron nightly mixed precision checkpoints test #27284

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6da11ea
more
Jun 4, 2026
3b194cc
more
Jun 4, 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
14 changes: 8 additions & 6 deletions python/sglang/srt/model_executor/model_runner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1419,14 +1419,16 @@ def load_model(self):
)

# TODO: Make sure all models have `quant_config` attribute, and all online quantization methods register which layers they actually quantize.
# TODO: Move this online-quantization reporting out of ModelRunner.
quantized_layers = getattr(
getattr(self.model, "quant_config", None), "quantized_layers", None
)
if (
hasattr(self.model, "quant_config")
and hasattr(self.model.quant_config, "quantized_layers")
and self.server_args.quantization is not None
self.server_args.quantization is not None
and isinstance(quantized_layers, tuple)
and len(quantized_layers) == 2
):
layer_types, quantized_layers_count = (
self.model.quant_config.quantized_layers
)
layer_types, quantized_layers_count = quantized_layers
logger.info(
f"Online {self.server_args.quantization} quantization: quantized {quantized_layers_count} layers of types: {layer_types}"
)
Expand Down
Loading