Emotion-classifier

I have build an emotion detection model which detects emotions from texts, By finetuning BERT transformer model(bert-base-uncased).

• Dataset used : Go Emotions dataset(Google)

• I have done Multi-label classification over here.

• Hosted on Hugging Face Spaces too

  HF space link : https://huggingface.co/spaces/nayadrika/emotion_detection_model

• Loss function used : BCEWithLogitsLoss

• Model is trained to recognize 28 emotions,

  such as :

#	Emotion	#	Emotion
1	admiration	15	fear
2	amusement	16	gratitude
3	anger	17	grief
4	annoyance	18	joy
5	approval	19	love
6	caring	20	nervousness
7	confusion	21	optimism
8	curiosity	22	pride
9	desire	23	realization
10	disappointment	24	relief
11	disapproval	25	remorse
12	disgust	26	sadness
13	embarrassment	27	surprise
14	excitement	28	neutral

📊 Evaluation Metrics

Metric	Score
Micro F1	0.606
Macro F1	0.466
Precision	0.592
Recall	0.620
Evaluation Loss	0.169