results_1

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

• Loss: 3.9021
• Accuracy: 0.5208
• Precision: 0.5264
• Recall: 0.5208
• F1: 0.5209

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-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 200
• num_epochs: 24

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.0903	1.0	120	1.1347	0.3333	0.4562	0.3333	0.2515
1.0362	2.0	240	1.1535	0.4125	0.4076	0.4125	0.3373
0.7914	3.0	360	1.2829	0.4417	0.4759	0.4417	0.4083
0.9578	4.0	480	1.2970	0.4542	0.4829	0.4542	0.4573
0.4566	5.0	600	1.8567	0.4708	0.5151	0.4708	0.4427
0.2706	6.0	720	2.2959	0.4875	0.5072	0.4875	0.4891
0.2104	7.0	840	2.3196	0.4958	0.4916	0.4958	0.4923
0.1874	8.0	960	2.8726	0.4667	0.4869	0.4667	0.4703
0.1814	9.0	1080	3.1062	0.4917	0.5477	0.4917	0.4825
0.1024	10.0	1200	3.3202	0.4792	0.4819	0.4792	0.4761
0.0608	11.0	1320	3.4888	0.5167	0.5326	0.5167	0.5180
0.0107	12.0	1440	3.5569	0.5042	0.5036	0.5042	0.5008
0.0003	13.0	1560	3.8422	0.5125	0.5316	0.5125	0.5070
0.0002	14.0	1680	3.8754	0.4958	0.5103	0.4958	0.4936
0.0002	15.0	1800	3.9021	0.5208	0.5264	0.5208	0.5209
0.0006	16.0	1920	4.2446	0.4833	0.4988	0.4833	0.4703
0.0001	17.0	2040	4.0457	0.4958	0.5034	0.4958	0.4945
0.0001	18.0	2160	4.1199	0.5167	0.5115	0.5167	0.5079
0.0001	19.0	2280	4.2421	0.5	0.5115	0.5	0.4939
0.0001	20.0	2400	4.3337	0.4958	0.5153	0.4958	0.4920
0.0001	21.0	2520	4.3489	0.4792	0.4921	0.4792	0.4756
0.0001	22.0	2640	4.3445	0.4833	0.4964	0.4833	0.4805
0.0001	23.0	2760	4.3644	0.4792	0.4921	0.4792	0.4756
0.0001	24.0	2880	4.3674	0.4833	0.4972	0.4833	0.4792

Framework versions

• Transformers 4.40.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1