Update README.md
Browse files
README.md
CHANGED
|
@@ -13,6 +13,12 @@ widget:
|
|
| 13 |
- text: 'You can email me at [email protected] or call directly on 9999988888. The point of contact would be my manager Avisekh Raj'
|
| 14 |
example_title: Email 3
|
| 15 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
|
| 17 |
|Class| Precision | Recall | f1 |
|
| 18 |
|-----|----------|:-------------:|------:|
|
|
@@ -20,4 +26,19 @@ widget:
|
|
| 20 |
| 1 | 0.98 | 1.00 | 0.91 |
|
| 21 |
| 2 | 0.95 | 0.89 | 0.92 |
|
| 22 |
| 3 | 0.8 | 0.88 | 0.84 |
|
| 23 |
-
| macro-avg | 0.93 | 0.94 | 0.94 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
- text: 'You can email me at [email protected] or call directly on 9999988888. The point of contact would be my manager Avisekh Raj'
|
| 14 |
example_title: Email 3
|
| 15 |
---
|
| 16 |
+
Overview:
|
| 17 |
+
|
| 18 |
+
The Model is fine-tuned for 3 class + "0" class.
|
| 19 |
+
The Dataset is custom annotated and contains 400 texts and the model was trained on the split of 0.76, 0.12, and 0.12.
|
| 20 |
+
|
| 21 |
+
The validation classification report is as follows:
|
| 22 |
|
| 23 |
|Class| Precision | Recall | f1 |
|
| 24 |
|-----|----------|:-------------:|------:|
|
|
|
|
| 26 |
| 1 | 0.98 | 1.00 | 0.91 |
|
| 27 |
| 2 | 0.95 | 0.89 | 0.92 |
|
| 28 |
| 3 | 0.8 | 0.88 | 0.84 |
|
| 29 |
+
| macro-avg | 0.93 | 0.94 | 0.94 |
|
| 30 |
+
|
| 31 |
+
The test classification report is as follows:
|
| 32 |
+
|
| 33 |
+
|Class| Precision | Recall | f1 |
|
| 34 |
+
|-----|----------|:-------------:|------:|
|
| 35 |
+
| 0 | 1.00 | 1.00 | 1.00 |
|
| 36 |
+
| 1 | 0.98 | 1.00 | 0.99 |
|
| 37 |
+
| 2 | 0.66 | 0.97 | 0.79 |
|
| 38 |
+
| 3 | 0.84 | 0.78 | 0.81 |
|
| 39 |
+
| macro-avg | 0.87 | 0.94 | 0.90 |
|
| 40 |
+
|
| 41 |
+
Possible future direction:
|
| 42 |
+
|
| 43 |
+
1. Clean data to a good enough format as much as possible.
|
| 44 |
+
2. Increase the data as much as possible. (Make sure to have data that is seen in real-usecase.)
|