Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

About the number of key frames used

See original GitHub issue

Hi! Thank you for your code and sorry to bother again.

I checked your log and found that in AVA logs, the key frames used are:

[INFO: ava.py:  101]: Number of unique boxes: 666191
[INFO: ava.py:  102]: Number of annotations: 1607900
[INFO: ava.py:  123]: 193301 keyframes used.

However, when I ran your code (ava_train_v2.1.csv and ava_train_predicted_boxes.csv are downloaded accroding to your instructions), my output log looks like:

[INFO: ava.py:  101]: Number of unique boxes: 664550
[INFO: ava.py:  102]: Number of annotations: 1603805
[INFO: ava.py:  123]: 193296 keyframes used.

I also took out the code snippet of load_boxes_and_labels() in ava.py and ran it, getting the same results.

I’m wondering where the difference comes from. Thanks a lot!

Issue Analytics

State:
Created 4 years ago
Comments:11

Top GitHub Comments

2reactions

wjchaoGitcommented, Aug 30, 2019

Hi, @chenz97 @chaoyuaw Issue #16 may lead to this problem.

When I ran the original code with the new version csv file, the number of key frames used is 193296. After modifying the code, the number of key frames used is 193301.

1reaction

chenz97commented, Aug 30, 2019

Hi @wjchaoGit , thank you very much!! That’s it!

Hi @chaoyuaw , yes, the issue gets fixed! Thank you for your help with great patience!!