Can't get comparable accuracy
Nobody has claimed this yet.
- Dominant language
- Jupyter Notebook
- Stars
- 9.8k
- Forks
- 1.6k
- PR merge metrics
- No merged PRs in 30d
Description
Bug Report
Which model does this pertain to?
https://github.com/onnx/models/tree/main/vision/object_detection_segmentation/retinanet
https://github.com/onnx/models/tree/main/text/machine_comprehension/roberta
Describe the bug
For retinanet, with preprocess and postprocess code in readme, I can't get mAP 0.376.
For roberta, I use RoBERTa-SequenceClassification model and MRPC dataset and calculate GLUE, its drop a lot (it is 90.2 in readme)
Reproduction instructions
System Information
OS Platform and Distribution (e.g. Linux Ubuntu 16.04):
ONNX version (e.g. 1.6): 1.10.0
Backend/Runtime version (e.g. ONNX Runtime 1.1, PyTorch 1.2): ORT 1.11.0
Notes
Any additional information
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with the RetinaNet and RoBERTa model READMEs linked in the issue, including their preprocessing, postprocessing, and reported evaluation values. Reproduce the RetinaNet mAP and RoBERTa MRPC/GLUE results using ONNX 1.10.0 and ORT 1.11.0, then document the inputs, commands, and point where each result diverges.
Written by the indexing model from the issue text.
Assessment
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100