STS-conventional-Fine-Tuning-Capstone-roberta-base-filtered-200

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

• Loss: 2.1218
• Accuracy: 0.7322

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: 3e-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
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	113	0.7870	0.6704
No log	2.0	226	0.7207	0.6779
No log	3.0	339	0.7853	0.7022
No log	4.0	452	0.8490	0.6742
0.513	5.0	565	1.0300	0.6835
0.513	6.0	678	1.1645	0.7060
0.513	7.0	791	1.4119	0.7210
0.513	8.0	904	1.5641	0.7154
0.1552	9.0	1017	1.7410	0.6966
0.1552	10.0	1130	1.8357	0.7228
0.1552	11.0	1243	1.9442	0.7116
0.1552	12.0	1356	1.9521	0.7266
0.1552	13.0	1469	2.0899	0.7172
0.0481	14.0	1582	2.1138	0.7210
0.0481	15.0	1695	2.1218	0.7322

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2