finetuning-emotion-roberta

This model is a fine-tuned version of roberta-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3262
• Accuracy: 0.9365
• F1: 0.9366

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	F1
No log	1.0	250	0.2507	0.9145	0.9158
0.4547	2.0	500	0.1703	0.9305	0.9293
0.4547	3.0	750	0.1722	0.9335	0.9345
0.1329	4.0	1000	0.1377	0.939	0.9382
0.1329	5.0	1250	0.1443	0.941	0.9411
0.0979	6.0	1500	0.1355	0.936	0.9365
0.0979	7.0	1750	0.1581	0.94	0.9394
0.0788	8.0	2000	0.1680	0.9375	0.9378
0.0788	9.0	2250	0.1876	0.9345	0.9342
0.0593	10.0	2500	0.2207	0.9335	0.9342
0.0593	11.0	2750	0.2065	0.937	0.9375
0.0463	12.0	3000	0.2185	0.939	0.9390
0.0463	13.0	3250	0.2239	0.938	0.9380
0.0354	14.0	3500	0.2555	0.932	0.9320
0.0354	15.0	3750	0.3019	0.933	0.9330
0.0241	16.0	4000	0.3129	0.935	0.9351
0.0241	17.0	4250	0.3152	0.939	0.9387
0.0202	18.0	4500	0.3228	0.9345	0.9347
0.0202	19.0	4750	0.3224	0.937	0.9371
0.0148	20.0	5000	0.3262	0.9365	0.9366

Framework versions

• Transformers 4.44.2
• Pytorch 2.5.0+cu121
• Datasets 3.1.0
• Tokenizers 0.19.1