Robustness low quality phone¶
CNN14 |
w2v2-b |
hubert-b |
axlstm |
|
---|---|---|---|---|
Overall Score |
25.0% passed tests (1 passed / 3 failed). |
75.0% passed tests (3 passed / 1 failed). |
25.0% passed tests (1 passed / 3 failed). |
0.0% passed tests (0 passed / 4 failed). |
Change Ccc Low Quality Phone¶
Data |
Change CCC Low Quality Phone |
|||
---|---|---|---|---|
CNN14 |
w2v2-b |
hubert-b |
axlstm |
|
iemocap-2.3.0-emotion.dimensions.test.gold_standard |
0.02 |
-0.06 |
-0.10 |
-0.11 |
msppodcast-2.6.1-emotion.dimensions.test-1.gold_standard |
-0.07 |
-0.03 |
-0.08 |
-0.16 |
mean |
-0.03 |
-0.04 |
-0.09 |
-0.14 |
Percentage Unchanged Predictions Low Quality Phone¶
Data |
Percentage Unchanged Predictions Low Quality Phone |
|||
---|---|---|---|---|
CNN14 |
w2v2-b |
hubert-b |
axlstm |
|
iemocap-2.3.0-emotion.dimensions.test.gold_standard |
0.27 |
0.52 |
0.52 |
0.45 |
msppodcast-2.6.1-emotion.dimensions.test-1.gold_standard |
0.37 |
0.51 |
0.48 |
0.29 |
mean |
0.32 |
0.52 |
0.50 |
0.37 |
Visualization¶
Difference of predictions for original audio and low quality phone audio. The allowed prediction difference \(\delta < 0.05\) is highlighted in green in the upper plot. The lower plot shows the distributions of the two predictions.
CNN14 |
w2v2-b |
hubert-b |
axlstm |
---|---|---|---|