Decathlon / Decathlon/decavision
ResNet support
- Dominant language
- Python
- Stars
- 4
- Forks
- 3
- PR merge metrics
- No merged PRs in 30d
Description
Resnet is supported by decavision according to the docstring of the `ImageClassifier` class and has seen in the code [here](https://github.com/Decathlon/decavision/blob/master/decavision/model_training/tfrecords_image_classifier.py#L254).
However, Resnet seems to be missing from the dict of [input_dims](https://github.com/Decathlon/decavision/blob/master/decavision/model_training/tfrecords_image_classifier.py#L100)
Contributor guide
Research direction
Read decavision/model_training/tfrecords_image_classifier.py, especially the input_dims dictionary around line 100 and the ResNet-related path around line 254. Confirm how ResNet is selected and what dimensions it requires; done means the supported ResNet model has a matching input_dims entry and the existing classifier path can use it.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, tensorflow
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Stale
- Clarity
- Clearly specified
- Newbie friendliness
- 50/100