Skip to content

Conversation

@idantene
Copy link

@idantene idantene commented Aug 1, 2025

In splitting an onnx file with the current code, fill_input_encodings_of_split can fail because it assumes encodings["X_params"] is a dictionary, which is no longer the case with the new AIMET spec for version 1.0.0.
Crucially, this is fixed in the split_onnx_by_names method, but not in the _load_encodings one, so this is a failure point.

@bhushan23
Copy link
Contributor

Hi @idantene thank you for putting up this PR, changes looks great.

We will take this up and merge internally.
Minor change, we can gate this function upon encoding_version so we can support all three versions that we have to make it robust.

@idantene
Copy link
Author

idantene commented Aug 6, 2025

Sure, as per the docs I'm only aware of 2 versions (pre-1.0.0 and 1.0.0), so it's interesting to learn you have three versions out.

@qaihm-bot qaihm-bot force-pushed the main branch 27 times, most recently from ae856a1 to 1d1790e Compare September 19, 2025 23:25
@qaihm-bot qaihm-bot force-pushed the main branch 4 times, most recently from 8d5d11c to 09ece6a Compare September 22, 2025 16:23
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants