Skip to content

fix: max_positions reads the text config (was 0 for Gemma 4) #95

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
AdvayMonga merged 1 commit into from
Oct 5, 2026
Merged

fix: max_positions reads the text config (was 0 for Gemma 4) #95

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
2 changes: 1 addition & 1 deletion src/inference_server/backends/custom_torch_backend.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -264,7 +264,7 @@ def load_model(self, model_name: str) -> None:
self.model = model_cls.from_safetensors(model_name, device=self.device)
else:
self.model = model_cls.from_hf(model_name, dtype=torch.bfloat16).to(self.device).eval()
self.max_positions = int(getattr(AutoConfig.from_pretrained(model_name), "max_position_embeddings", 0) or 0)
self.max_positions = int(getattr(AutoConfig.from_pretrained(model_name).get_text_config(), "max_position_embeddings", 0) or 0) # Gemma 4 keeps it on the text config
self._eos_ids = stop_token_ids(model_name, self.tokenizer)

# Weight-only int8: store Linear weights as int8 (per-channel), read 2× fewer bytes/step.
Expand Down
Loading