bert-dair-ai-emotion-testing

This model is a fine-tuned version of bert-base-cased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1413
• F1: 0.8670

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: 5e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1
1.4777	1.0	48	1.1115	0.6293
1.0781	2.0	96	0.7560	0.6536
0.7411	3.0	144	0.6211	0.7324
0.3429	4.0	192	0.4169	0.7486
0.4494	5.0	240	0.2302	0.7559
0.176	6.0	288	0.1959	0.8222
0.1838	7.0	336	0.1578	0.8647
0.1846	8.0	384	0.1451	0.8304
0.1379	9.0	432	0.1554	0.8647
0.0895	10.0	480	0.1418	0.8328
0.0151	11.0	528	0.1468	0.8304
0.0625	12.0	576	0.1630	0.8304
0.0397	13.0	624	0.1372	0.8304
0.0177	14.0	672	0.1359	0.8304
0.0062	15.0	720	0.1386	0.8328
0.0244	16.0	768	0.1298	0.8351
0.01	17.0	816	0.1369	0.8351
0.0094	18.0	864	0.1418	0.8670
0.0329	19.0	912	0.1400	0.8670
0.0791	20.0	960	0.1413	0.8670

Framework versions

• PEFT 0.13.2
• Transformers 4.48.0
• Pytorch 2.4.1+cu124
• Datasets 3.0.1
• Tokenizers 0.21.0