N_roberta_twitterfin_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: 0.8997
• Accuracy: 0.9062

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
0.5296	1.0	597	0.2984	0.8903
0.3008	2.0	1194	0.2405	0.9121
0.2473	3.0	1791	0.4214	0.8807
0.1821	4.0	2388	0.4822	0.9049
0.1559	5.0	2985	0.5408	0.9008
0.0825	6.0	3582	0.5540	0.9070
0.0745	7.0	4179	0.6724	0.9016
0.0498	8.0	4776	0.7321	0.8945
0.0502	9.0	5373	0.7188	0.8991
0.0501	10.0	5970	0.7382	0.9058
0.0299	11.0	6567	0.7618	0.9008
0.0263	12.0	7164	0.7383	0.9075
0.0191	13.0	7761	0.7731	0.9066
0.0133	14.0	8358	0.8335	0.9054
0.0149	15.0	8955	0.8198	0.9049
0.0109	16.0	9552	0.8299	0.9054
0.0155	17.0	10149	0.8402	0.9066
0.0027	18.0	10746	0.8514	0.9087
0.0024	19.0	11343	0.8878	0.9070
0.004	20.0	11940	0.8997	0.9062

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.13.3