distilbert-base-uncased-finetuned-emotion

This model is a fine-tuned version of distilbert-base-uncased on the emotion dataset. It achieves the following results on the evaluation set:

• Loss: 0.7264
• Accuracy: 0.895
• Balanced accuracy: 0.8746
• F1: 0.8961

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Balanced accuracy	F1
0.001	1.0	25	0.7713	0.89	0.8807	0.8915
0.0069	2.0	50	0.7734	0.905	0.8906	0.9070
0.0019	3.0	75	0.8670	0.88	0.8749	0.8819
0.0012	4.0	100	0.7387	0.895	0.8806	0.8953
0.0002	5.0	125	0.7841	0.885	0.8649	0.8858
0.0002	6.0	150	0.7415	0.9	0.8753	0.9001
0.0002	7.0	175	0.7378	0.895	0.8719	0.8955
0.0002	8.0	200	0.7452	0.89	0.8711	0.8910
0.0002	9.0	225	0.7555	0.89	0.8787	0.8908
0.0001	10.0	250	0.7541	0.895	0.8822	0.8959
0.0001	11.0	275	0.7536	0.9	0.8857	0.9009
0.0001	12.0	300	0.7530	0.9	0.8857	0.9009
0.0001	13.0	325	0.7542	0.9	0.8857	0.9009
0.0001	14.0	350	0.7532	0.895	0.8746	0.8957
0.0002	15.0	375	0.8554	0.88	0.8424	0.8803
0.0001	16.0	400	0.7700	0.9	0.8867	0.9011
0.0001	17.0	425	0.7302	0.895	0.8746	0.8961
0.0001	18.0	450	0.7304	0.895	0.8746	0.8961
0.0001	19.0	475	0.7284	0.895	0.8746	0.8961
0.0001	20.0	500	0.7264	0.895	0.8746	0.8961

Framework versions

• Transformers 4.38.1
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2