non_green_as_train_contextroberta-large_20e

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

• Loss: 0.3214
• Val Accuracy: 0.9779
• Val Precision: 0.6893
• Val Recall: 0.7568
• Val F1: 0.7215

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: 5e-06
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Val Accuracy	Val Precision	Val Recall	Val F1
0.0601	1.0	7739	0.0767	0.9763	0.6646	0.7518	0.7055
0.0493	2.0	15478	0.0995	0.9785	0.7181	0.7094	0.7137
0.0305	3.0	23217	0.1216	0.9765	0.6670	0.7578	0.7095
0.0196	4.0	30956	0.1275	0.9786	0.7066	0.7437	0.7247
0.0161	5.0	38695	0.1521	0.9768	0.7164	0.6398	0.6759
0.0141	6.0	46434	0.1643	0.9785	0.7103	0.7275	0.7188
0.007	7.0	54173	0.1660	0.9769	0.6739	0.7528	0.7112
0.0052	8.0	61912	0.1855	0.9783	0.7036	0.7376	0.7202
0.0048	9.0	69651	0.1845	0.9781	0.7042	0.7255	0.7147
0.0031	10.0	77390	0.2165	0.9782	0.7225	0.6882	0.7049
0.0036	11.0	85129	0.2271	0.9783	0.7223	0.6902	0.7059
0.0029	12.0	92868	0.2345	0.9770	0.6887	0.7144	0.7013
0.0015	13.0	100607	0.2636	0.9781	0.7307	0.6680	0.6979
0.0045	14.0	108346	0.2493	0.9781	0.6846	0.7820	0.7301
0.0005	15.0	116085	0.2563	0.9774	0.6789	0.7639	0.7189
0.0007	16.0	123824	0.2856	0.9784	0.7193	0.7033	0.7112
0.0	17.0	131563	0.2809	0.9782	0.7136	0.7064	0.7099
0.0	18.0	139302	0.3033	0.9781	0.6957	0.7497	0.7217
0.0	19.0	147041	0.3207	0.9782	0.6909	0.7669	0.7269
0.0	20.0	154780	0.3214	0.9779	0.6893	0.7568	0.7215

Framework versions

• Transformers 4.38.2
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.15.2