Update README.md
Browse files
README.md
CHANGED
|
@@ -38,9 +38,6 @@ Using a fixed threshold of 0.5 to convert the scores to binary predictions for e
|
|
| 38 |
- Recall: 0.250
|
| 39 |
- F1: 0.303
|
| 40 |
|
| 41 |
-
Simple mean of labels: {'precision': 0.445, 'recall': 0.476, 'f1': 0.449}
|
| 42 |
-
Weighted average (using support): {'precision': 0.472, 'recall': 0.582, 'f1': 0.514}
|
| 43 |
-
|
| 44 |
Optimising the threshold per label to optimise the F1 metric, the metrics (evaluated on the go_emotions test split) are:
|
| 45 |
|
| 46 |
- Precision: 0.445
|
|
@@ -57,6 +54,39 @@ Weighted by the relative support of each label in the dataset, this is:
|
|
| 57 |
|
| 58 |
This is a multi-label, multi-class dataset, so each label is effectively a separate binary classification and metrics are better measured per label.
|
| 59 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 60 |
Using a fixed threshold of 0.5 to convert the scores to binary predictions for each label, the metrics (evaluated on the go_emotions test split) are:
|
| 61 |
|
| 62 |
| | f1 | precision | recall | support | threshold |
|
|
@@ -90,4 +120,8 @@ Using a fixed threshold of 0.5 to convert the scores to binary predictions for e
|
|
| 90 |
| surprise | 0.142 | 0.786 | 0.078 | 141 | 0.5 |
|
| 91 |
| neutral | 0.547 | 0.644 | 0.475 | 1787 | 0.5 |
|
| 92 |
|
|
|
|
| 93 |
|
|
|
|
|
|
|
|
|
|
|
|
| 38 |
- Recall: 0.250
|
| 39 |
- F1: 0.303
|
| 40 |
|
|
|
|
|
|
|
|
|
|
| 41 |
Optimising the threshold per label to optimise the F1 metric, the metrics (evaluated on the go_emotions test split) are:
|
| 42 |
|
| 43 |
- Precision: 0.445
|
|
|
|
| 54 |
|
| 55 |
This is a multi-label, multi-class dataset, so each label is effectively a separate binary classification and metrics are better measured per label.
|
| 56 |
|
| 57 |
+
Optimising the threshold per label to optimise the F1 metric, the metrics (evaluated on the go_emotions test split) are:
|
| 58 |
+
|
| 59 |
+
| | f1 | precision | recall | support | threshold |
|
| 60 |
+
| -------------- | ----- | --------- | ------ | ------- | --------- |
|
| 61 |
+
| admiration | 0.583 | 0.574 | 0.593 | 504 | 0.30 |
|
| 62 |
+
| amusement | 0.668 | 0.722 | 0.621 | 264 | 0.25 |
|
| 63 |
+
| anger | 0.350 | 0.309 | 0.404 | 198 | 0.15 |
|
| 64 |
+
| annoyance | 0.299 | 0.318 | 0.281 | 320 | 0.20 |
|
| 65 |
+
| approval | 0.338 | 0.281 | 0.425 | 351 | 0.15 |
|
| 66 |
+
| caring | 0.321 | 0.323 | 0.319 | 135 | 0.20 |
|
| 67 |
+
| confusion | 0.384 | 0.313 | 0.497 | 153 | 0.15 |
|
| 68 |
+
| curiosity | 0.467 | 0.432 | 0.507 | 284 | 0.20 |
|
| 69 |
+
| desire | 0.426 | 0.381 | 0.482 | 83 | 0.20 |
|
| 70 |
+
| disappointment | 0.210 | 0.147 | 0.364 | 151 | 0.10 |
|
| 71 |
+
| disapproval | 0.366 | 0.288 | 0.502 | 267 | 0.15 |
|
| 72 |
+
| disgust | 0.416 | 0.409 | 0.423 | 123 | 0.20 |
|
| 73 |
+
| embarrassment | 0.370 | 0.341 | 0.405 | 37 | 0.30 |
|
| 74 |
+
| excitement | 0.313 | 0.368 | 0.272 | 103 | 0.25 |
|
| 75 |
+
| fear | 0.615 | 0.677 | 0.564 | 78 | 0.40 |
|
| 76 |
+
| gratitude | 0.828 | 0.810 | 0.847 | 352 | 0.25 |
|
| 77 |
+
| grief | 0.545 | 0.600 | 0.500 | 6 | 0.85 |
|
| 78 |
+
| joy | 0.455 | 0.429 | 0.484 | 161 | 0.20 |
|
| 79 |
+
| love | 0.642 | 0.673 | 0.613 | 238 | 0.30 |
|
| 80 |
+
| nervousness | 0.350 | 0.412 | 0.304 | 23 | 0.60 |
|
| 81 |
+
| optimism | 0.439 | 0.417 | 0.462 | 186 | 0.20 |
|
| 82 |
+
| pride | 0.480 | 0.667 | 0.375 | 16 | 0.70 |
|
| 83 |
+
| realization | 0.232 | 0.191 | 0.297 | 145 | 0.10 |
|
| 84 |
+
| relief | 0.353 | 0.500 | 0.273 | 11 | 0.50 |
|
| 85 |
+
| remorse | 0.643 | 0.529 | 0.821 | 56 | 0.20 |
|
| 86 |
+
| sadness | 0.526 | 0.497 | 0.558 | 156 | 0.20 |
|
| 87 |
+
| surprise | 0.329 | 0.318 | 0.340 | 141 | 0.15 |
|
| 88 |
+
| neutral | 0.634 | 0.528 | 0.794 | 1787 | 0.30 |
|
| 89 |
+
|
| 90 |
Using a fixed threshold of 0.5 to convert the scores to binary predictions for each label, the metrics (evaluated on the go_emotions test split) are:
|
| 91 |
|
| 92 |
| | f1 | precision | recall | support | threshold |
|
|
|
|
| 120 |
| surprise | 0.142 | 0.786 | 0.078 | 141 | 0.5 |
|
| 121 |
| neutral | 0.547 | 0.644 | 0.475 | 1787 | 0.5 |
|
| 122 |
|
| 123 |
+
### Use with ONNXRuntime
|
| 124 |
|
| 125 |
+
```python
|
| 126 |
+
pass
|
| 127 |
+
```
|