roberta_sst2_padding50model

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

• Loss: 0.5107
• Accuracy: 0.9462

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: 16
• eval_batch_size: 16
• 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
No log	1.0	433	0.1735	0.9319
0.327	2.0	866	0.2500	0.9336
0.1893	3.0	1299	0.2987	0.9407
0.1229	4.0	1732	0.3376	0.9418
0.0753	5.0	2165	0.3283	0.9484
0.0496	6.0	2598	0.5720	0.9116
0.0349	7.0	3031	0.4278	0.9363
0.0349	8.0	3464	0.4501	0.9379
0.0254	9.0	3897	0.4728	0.9374
0.0217	10.0	4330	0.4662	0.9368
0.0171	11.0	4763	0.4622	0.9418
0.0082	12.0	5196	0.4804	0.9429
0.0094	13.0	5629	0.4789	0.9445
0.0047	14.0	6062	0.5459	0.9423
0.0047	15.0	6495	0.4672	0.9434
0.009	16.0	6928	0.5178	0.9445
0.0021	17.0	7361	0.5107	0.9467
0.0042	18.0	7794	0.5101	0.9445
0.0053	19.0	8227	0.5043	0.9462
0.0017	20.0	8660	0.5107	0.9462

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.1
• Datasets 2.12.0
• Tokenizers 0.13.3