Allow dynamic batch-sizes when exporting to ONNX, instead if fixed input shapes #1347
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
When exporting to ONNX the models input shape is fixed to a batch-size of 1. This only allows to feed single images into the resuting anomaly model.
Torch.onnx allows to define dynamic axes when exporting the model, which enables dynamic batch-sizes.
Changes
Describe the changes you made
Checklist
Ensure that you followed the following