- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Everyone,
we are facing few issues while setting up the virtual env for Custom object detection model training related to the train.py script path. we have tried setting this up with master and few other older commits. Attaching the details over here.
Link followed for installation and Setup:(Custom Object Detection Training Example)
https://github.com/openvinotoolkit/training_extensions/tree/4c85176f611fad2f9c30e11be61a548ba7534db3
Link to Log File:
Regards,
Akshya
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello Akshya,
Greetings to you.
I replicate on my side and get a similar error. We are investigating this and will get back to you soon.
Sincerely,
Zulkifli
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Akshya,
We are able to run the training by following these steps:
- Refer to the training_extention documentation, and follow step 1 to 5.
- Once completed, deactivate the virtual environment.
- Refer to the custom object detection documentation, and follow step 1 to 7. You will be able to see (detection) as your new environment.
If you encounter this error:
ModuleNotFoundError: No module named 'mmcv._ext'
Follow these steps to resolve it:
1. Pip3 uninstall mmcv mmcv-full
2. git clone https://github.com/open-mmlab/mmcv.git
3. cd mmcv
4. git checkout v1.2.7
5. MMCV_WITH_OPS=1 python3 -m pip install -e .
Sincerely,
Zulkifli
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello Akshya,
This thread will no longer be monitored since we have provided a solution. If you need any additional information from Intel, please submit a new question.
Sincerely,
Zulkifli
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page