finetuning-emotion-model-roberta

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

• Loss: 0.3117
• Accuracy: 0.9405
• F1: 0.9403

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	0.2637	0.9135	0.9146
0.4901	2.0	500	0.1735	0.934	0.9341
0.4901	3.0	750	0.1678	0.9375	0.9381
0.1058	4.0	1000	0.1603	0.9385	0.9383
0.1058	5.0	1250	0.1902	0.9345	0.9346
0.0684	6.0	1500	0.1899	0.9365	0.9367
0.0684	7.0	1750	0.2289	0.932	0.9313
0.0406	8.0	2000	0.2291	0.936	0.9361
0.0406	9.0	2250	0.2438	0.935	0.9353
0.0254	10.0	2500	0.2639	0.9375	0.9375
0.0254	11.0	2750	0.2863	0.9365	0.9365
0.017	12.0	3000	0.2814	0.938	0.9379
0.017	13.0	3250	0.2972	0.9355	0.9355
0.012	14.0	3500	0.2874	0.9355	0.9356
0.012	15.0	3750	0.2996	0.9405	0.9403
0.0098	16.0	4000	0.3110	0.9405	0.9403
0.0098	17.0	4250	0.3173	0.9365	0.9364
0.0082	18.0	4500	0.3106	0.9405	0.9404
0.0082	19.0	4750	0.3100	0.94	0.9398
0.006	20.0	5000	0.3117	0.9405	0.9403

Framework versions

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