Update README.md
Browse files
README.md
CHANGED
|
@@ -10,15 +10,31 @@ pipeline_tag: text-classification
|
|
| 10 |
library_name: transformers
|
| 11 |
---
|
| 12 |
|
| 13 |
-
|
| 14 |
#### Overview
|
| 15 |
|
|
|
|
| 16 |
Model trained from [mobileBert](https://huggingface.co/google/mobilebert-uncased) on the [go_emotions](https://huggingface.co/datasets/google-research-datasets/go_emotions) dataset for multi-label classification.
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
|
| 18 |
#### Dataset used for the model
|
| 19 |
|
| 20 |
[go_emotions](https://huggingface.co/datasets/google-research-datasets/go_emotions) is based on Reddit data and has 28 labels. It is a multi-label dataset where one or multiple labels may apply for any given input text, hence this model is a multi-label classification model with 28 'probability' float outputs for any given input text. Typically a threshold of 0.5 is applied to the probabilities for the prediction for each label.
|
| 21 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
#### How the model was created
|
| 23 |
|
| 24 |
The model was trained using `AutoModelForSequenceClassification.from_pretrained` with `problem_type="multi_label_classification"` for 3 epochs with a learning rate of 2e-5 and weight decay of 0.01.
|
|
|
|
| 10 |
library_name: transformers
|
| 11 |
---
|
| 12 |
|
|
|
|
| 13 |
#### Overview
|
| 14 |
|
| 15 |
+
|
| 16 |
Model trained from [mobileBert](https://huggingface.co/google/mobilebert-uncased) on the [go_emotions](https://huggingface.co/datasets/google-research-datasets/go_emotions) dataset for multi-label classification.
|
| 17 |
+
<div align="center">
|
| 18 |
+
<a href=https://github.com/04AR/Senti target="_blank"><img src=https://img.shields.io/badge/Code-black.svg?logo=github height=22px></a>
|
| 19 |
+
<a href=https://huggingface.co/AR04/Senti target="_blank"><img src=https://img.shields.io/badge/%F0%9F%A4%97%20Models-d96902.svg height=22px></a>
|
| 20 |
+
</div>
|
| 21 |
|
| 22 |
#### Dataset used for the model
|
| 23 |
|
| 24 |
[go_emotions](https://huggingface.co/datasets/google-research-datasets/go_emotions) is based on Reddit data and has 28 labels. It is a multi-label dataset where one or multiple labels may apply for any given input text, hence this model is a multi-label classification model with 28 'probability' float outputs for any given input text. Typically a threshold of 0.5 is applied to the probabilities for the prediction for each label.
|
| 25 |
|
| 26 |
+
## 🏷️ Emotion Labels
|
| 27 |
+
|
| 28 |
+
| ID | Emotion | ID | Emotion | ID | Emotion | ID | Emotion |
|
| 29 |
+
|-----|----------------|-----|----------------|-----|----------------|-----|----------------|
|
| 30 |
+
| 0 | admiration | 1 | amusement | 2 | anger | 3 | annoyance |
|
| 31 |
+
| 4 | approval | 5 | caring | 6 | confusion | 7 | curiosity |
|
| 32 |
+
| 8 | desire | 9 | disappointment | 10 | disapproval | 11 | disgust |
|
| 33 |
+
| 12 | embarrassment | 13 | excitement | 14 | fear | 15 | gratitude |
|
| 34 |
+
| 16 | grief | 17 | joy | 18 | love | 19 | nervousness |
|
| 35 |
+
| 20 | optimism | 21 | pride | 22 | realization | 23 | relief |
|
| 36 |
+
| 24 | remorse | 25 | sadness | 26 | surprise | 27 | neutral |
|
| 37 |
+
|
| 38 |
#### How the model was created
|
| 39 |
|
| 40 |
The model was trained using `AutoModelForSequenceClassification.from_pretrained` with `problem_type="multi_label_classification"` for 3 epochs with a learning rate of 2e-5 and weight decay of 0.01.
|