search

zyh-uaiaaaa / Erasing-Attention-Consistency

Official implementation of the ECCV2022 paper: Learn From All: Erasing Attention Consistency for Noisy Label Facial Expression Recognition
77 stars 15 forks source link

Visualization of the classification loss values problem #23

Open amazedan opened 6 months ago

amazedan commented 6 months ago

Thank you for your great works! I'm not quite clear on how the visualization map in Section 4.7 were generated(Fig. 5). Could you please share the relevant repo links or code references? I appreciate your response to my questions.

zyh-uaiaaaa commented 6 months ago

Hi Amazedan,

Firstly, you can assess the performance of the trained model by evaluating it on all test samples and obtaining the loss values for each sample. Afterward, you can utilize matplotlib to create a histogram depicting the distribution of the loss values.

amazedan commented 6 months ago

Thank you! I think I understand how to draw now.

amazedan commented 4 months ago

Hello, author. I'd like to confirm with you: are we calculating the loss values on the test set? If so, how do we differentiate between noisy and clean samples?