TLT V2.0 Classification

Was self resolved.
I changed from yuw-v2, which was updated 3 days ago, to v2.0_py3 and it was successful.
Thank you.

Use jupyter.

Follow this URL.

3.Invoke the jupyter notebook using the following command
jupyter notebook --ip 0.0.0.0 --port 8888 --allow-root

4.Open an internet browser and type in the following URL to start running the notebooks when running on a local machine.
http://0.0.0.0:8888

Click on the maskrcnn.ipynb file as shown in the picture and run it sequentially.

@h12sw06,
Any problem on your side? Please create a new forum topic if needed.

Resizing image by 1/2 helped on frequent crashes. Still has exits with none descript error messages that appear to be actionable but much longer runtime.

When resizing the image by 1/2 does that also scale the mask/outline annotation by 1/2? Assuming work specific to mask rcnn was done on image resizing to scale outline annotations.

Results after 300k+ steps are reasonable for people and zebras but the rest not great.

Need to go back and compare results using the mask rcnn model before returning/transfer learning.

Thanks again for your help

@swillis,
If needed, could you please create a new forum topic and describe more details(via log, step, html format jupyter notebook, etc) about the new problem?
I’m asking this because there are several different questions/issues during this topic.

Will do. The forum software because I am a new user would accept a post then reject it because I was a new user or it thought it was spam. Behavior was very strange and had trouble figuring out what was accepted, rejected, accepted a couple days later.