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: 0.1910
• Accuracy: 0.9385
• F1: 0.9385

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: 256
• eval_batch_size: 256
• 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	F1
No log	1.0	63	0.3110	0.907	0.9066
No log	2.0	126	0.2093	0.9255	0.9264
No log	3.0	189	0.1683	0.931	0.9312
0.2667	4.0	252	0.1564	0.932	0.9319
0.2667	5.0	315	0.1541	0.9325	0.9328
0.2667	6.0	378	0.1577	0.9375	0.9378
0.2667	7.0	441	0.1547	0.9355	0.9357
0.0894	8.0	504	0.1528	0.9385	0.9386
0.0894	9.0	567	0.1630	0.9395	0.9394
0.0894	10.0	630	0.1745	0.9425	0.9427
0.0894	11.0	693	0.1635	0.9385	0.9385
0.0567	12.0	756	0.1706	0.938	0.9381
0.0567	13.0	819	0.1740	0.941	0.9413
0.0567	14.0	882	0.1766	0.94	0.9403
0.0567	15.0	945	0.1832	0.938	0.9382
0.0397	16.0	1008	0.1871	0.9385	0.9388
0.0397	17.0	1071	0.1889	0.938	0.9382
0.0397	18.0	1134	0.1908	0.935	0.9354
0.0397	19.0	1197	0.1907	0.94	0.9399
0.0284	20.0	1260	0.1910	0.9385	0.9385

Framework versions

• Transformers 4.38.1
• Pytorch 2.2.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2