classifier_roberta

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

• Loss: 0.6933
• Accuracy: 0.4392
• Precision: 0.4392
• Recall: 1.0

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: 0.001
• train_batch_size: 16
• eval_batch_size: 16
• seed: 0
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: polynomial
• num_epochs: 32

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall
No log	1.0	358	0.8518	0.4392	0.4392	1.0
0.7654	2.0	716	0.7826	0.4392	0.4392	1.0
0.7448	3.0	1074	0.8327	0.4392	0.4392	1.0
0.7448	4.0	1432	0.7101	0.5608	0.0	0.0
0.7469	5.0	1790	0.6857	0.5608	0.0	0.0
0.758	6.0	2148	0.6858	0.5608	0.0	0.0
0.7646	7.0	2506	0.7054	0.4392	0.4392	1.0
0.7646	8.0	2864	0.7196	0.4392	0.4392	1.0
0.7329	9.0	3222	0.6947	0.4392	0.4392	1.0
0.747	10.0	3580	0.7143	0.4392	0.4392	1.0
0.747	11.0	3938	0.6863	0.5608	0.0	0.0
0.7343	12.0	4296	0.6857	0.5608	0.0	0.0
0.7461	13.0	4654	0.7057	0.4392	0.4392	1.0
0.7279	14.0	5012	0.6893	0.5608	0.0	0.0
0.7279	15.0	5370	0.7015	0.4392	0.4392	1.0
0.735	16.0	5728	0.7138	0.4392	0.4392	1.0
0.73	17.0	6086	0.7042	0.5608	0.0	0.0
0.73	18.0	6444	0.7084	0.4392	0.4392	1.0
0.7299	19.0	6802	0.6978	0.4392	0.4392	1.0
0.7216	20.0	7160	0.6924	0.5608	0.0	0.0
0.7246	21.0	7518	0.7701	0.4392	0.4392	1.0
0.7246	22.0	7876	0.9114	0.4392	0.4392	1.0
0.7183	23.0	8234	0.8309	0.4392	0.4392	1.0
0.7158	24.0	8592	0.6875	0.5608	0.0	0.0
0.7158	25.0	8950	0.6875	0.5608	0.0	0.0
0.7112	26.0	9308	0.6857	0.5608	0.0	0.0
0.7097	27.0	9666	0.6913	0.5608	0.0	0.0
0.7076	28.0	10024	0.6996	0.4392	0.4392	1.0
0.7076	29.0	10382	0.7932	0.4392	0.4392	1.0
0.704	30.0	10740	0.6858	0.5608	0.0	0.0
0.7007	31.0	11098	0.6975	0.4392	0.4392	1.0
0.7007	32.0	11456	0.6933	0.4392	0.4392	1.0

Framework versions

• Transformers 4.44.2
• Pytorch 2.5.0+cu121
• Datasets 3.1.0
• Tokenizers 0.19.1