roberta-large-sst-2-64-13-30

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

• Loss: 0.8764
• Accuracy: 0.8828

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: 1.5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 5
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	4	0.7179	0.5
No log	2.0	8	0.6981	0.5312
0.717	3.0	12	0.6948	0.4688
0.717	4.0	16	0.7043	0.4453
0.6986	5.0	20	0.6971	0.4688
0.6986	6.0	24	0.7705	0.5156
0.6986	7.0	28	0.7463	0.625
0.6087	8.0	32	0.7016	0.6172
0.6087	9.0	36	0.5869	0.7656
0.5365	10.0	40	0.5156	0.8047
0.5365	11.0	44	0.4578	0.8203
0.5365	12.0	48	0.3511	0.9141
0.3599	13.0	52	0.3583	0.8828
0.3599	14.0	56	0.3759	0.8828
0.1271	15.0	60	0.4324	0.8906
0.1271	16.0	64	0.4806	0.8984
0.1271	17.0	68	0.5256	0.875
0.0516	18.0	72	0.6432	0.8906
0.0516	19.0	76	0.6976	0.875
0.0034	20.0	80	0.8148	0.875
0.0034	21.0	84	0.8401	0.8828
0.0034	22.0	88	0.8721	0.8828
0.0467	23.0	92	0.8001	0.8906
0.0467	24.0	96	0.8580	0.8828
0.0005	25.0	100	0.8849	0.875
0.0005	26.0	104	0.9024	0.875
0.0005	27.0	108	0.9125	0.875
0.0005	28.0	112	0.8686	0.8828
0.0005	29.0	116	0.8764	0.8828
0.0231	30.0	120	0.8764	0.8828

Framework versions

• Transformers 4.32.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.4.0
• Tokenizers 0.13.3