metadata

language: en
tags:
  - text-classification
  - pytorch
  - roberta
datasets:
  - go_emotions
license: mit
widget:
  - text: I am not feeling well today.

This model is trained for GoEmotions dataset which contains labeled 58k Reddit comments with 28 emotions

admiration, amusement, anger, annoyance, approval, caring, confusion, curiosity, desire, disappointment, disapproval, disgust, embarrassment, excitement, fear, gratitude, grief, joy, love, nervousness, optimism, pride, realization, relief, remorse, sadness, surprise + neutral

The training script is provided here: https://github.com/bsinghpratap/roberta_train_goEmotion

The model works well on most of the emotions except: 'desire', 'disgust', 'embarrassment', 'excitement', 'fear', 'grief', 'nervousness', 'pride', 'relief', 'remorse', 'surprise']
I'll try to fine-tune the model further and update here if RoBERTa achieves a better performance.

Some Training details:

Each text datapoint can have more than 1 label. Most of the training set had 1 label: Counter({1: 36308, 2: 6541, 3: 532, 4: 28, 5: 1})
So currently I just used the first label for each of the datapoint. Not ideal but it does a decent job.

Current Performance

============================================================
Emotion: admiration
============================================================
GoEmotions Paper: 0.65
RoBERTa: 0.62
Support: 504
============================================================
Emotion: amusement
============================================================
GoEmotions Paper: 0.80
RoBERTa: 0.78
Support: 252
============================================================
Emotion: anger
============================================================
GoEmotions Paper: 0.47
RoBERTa: 0.44
Support: 197
============================================================
Emotion: annoyance
============================================================
GoEmotions Paper: 0.34
RoBERTa: 0.22
Support: 286
============================================================
Emotion: approval
============================================================
GoEmotions Paper: 0.36 RoBERTa: 0.31 Support: 318

Emotion: caring

GoEmotions Paper: 0.39 RoBERTa: 0.24 Support: 114

Emotion: confusion

GoEmotions Paper: 0.37 RoBERTa: 0.29 Support: 139

Emotion: curiosity

GoEmotions Paper: 0.54 RoBERTa: 0.48 Support: 233

Emotion: disappointment

GoEmotions Paper: 0.28 RoBERTa: 0.18 Support: 127

Emotion: disapproval

GoEmotions Paper: 0.39 RoBERTa: 0.26 Support: 220

Emotion: gratitude

GoEmotions Paper: 0.86 RoBERTa: 0.84 Support: 288

Emotion: joy

GoEmotions Paper: 0.51 RoBERTa: 0.47 Support: 116

Emotion: love

GoEmotions Paper: 0.78 RoBERTa: 0.68 Support: 169

Emotion: neutral

GoEmotions Paper: 0.68 RoBERTa: 0.61 Support: 1606

Emotion: optimism

GoEmotions Paper: 0.51 RoBERTa: 0.52 Support: 120

Emotion: realization

GoEmotions Paper: 0.21 RoBERTa: 0.15 Support: 109

Emotion: sadness

GoEmotions Paper: 0.49 RoBERTa: 0.42 Support: 108