Skip to content

convert : fix gpt-oss #18008

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

Closed
ggerganov wants to merge 2 commits into from
+2 −1
Closed

convert : fix gpt-oss #18008

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
68670ae
convert : fix gpt-oss
ggerganov Dec 13, 2025
4b4264f
Update convert_hf_to_gguf.py
ggerganov Dec 13, 2025
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
3 changes: 2 additions & 1 deletion convert_hf_to_gguf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9612,11 +9612,12 @@
self.gguf_writer.add_sliding_window(self.hparams["sliding_window"])
self.gguf_writer.add_expert_feed_forward_length(self.hparams["intermediate_size"])

rope_scaling = self.hparams.get("rope_scaling") or {}
rope_scaling = self.hparams.get("rope_scaling") or self.hparams.get("rope_parameters") or {}
rope_type = rope_scaling.get("rope_type", rope_scaling.get("type"))
assert rope_type == "yarn", f"GPT-OSS only supports yarn rope scaling, got {rope_type}"
self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.YARN)
self.gguf_writer.add_rope_scaling_factor(rope_scaling["factor"])
self.gguf_writer.add_rope_freq_base(rope_scaling.get("rope_theta"))

Check failure on line 9620 in convert_hf_to_gguf.py

View workflow job for this annotation

GitHub Actions / pyright type-check 

Argument of type "Unknown | Any | None" cannot be assigned to parameter "value" of type "float" in function "add_rope_freq_base"
  Type "Unknown | Any | None" is not assignable to type "float"
    "None" is not assignable to "float" (reportArgumentType)

Check failure on line 9620 in convert_hf_to_gguf.py

View workflow job for this annotation

GitHub Actions / pyright type-check 

Argument of type "Unknown | Any | None" cannot be assigned to parameter "value" of type "float" in function "add_rope_freq_base"
  Type "Unknown | Any | None" is not assignable to type "float"
    "None" is not assignable to "float" (reportArgumentType)
Copy link
Collaborator

@CISC CISC Dec 13, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
self.gguf_writer.add_rope_freq_base(rope_scaling.get("rope_theta"))
self.gguf_writer.add_rope_freq_base(rope_scaling.get("rope_theta", hparams.get("rope_theta")))

So, they moved this in v5, we need to check both places.

self.gguf_writer.add_rope_scaling_orig_ctx_len(rope_scaling.get("original_max_position_embeddings", 4096))


Expand Down
Loading