results_2

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: 2.3742
• Accuracy: 0.4708
• Precision: 0.4847
• Recall: 0.4708
• F1: 0.4734

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
1.1286	1.0	120	1.1322	0.3458	0.5471	0.3458	0.2569
1.007	2.0	240	1.1356	0.4625	0.4700	0.4625	0.4234
0.8347	3.0	360	1.2379	0.4292	0.4550	0.4292	0.3851
0.7291	4.0	480	1.4182	0.4583	0.4778	0.4583	0.4575
0.3339	5.0	600	1.9020	0.4792	0.4986	0.4792	0.4631
0.1862	6.0	720	2.3742	0.4708	0.4847	0.4708	0.4734
0.2639	7.0	840	2.9896	0.4542	0.4583	0.4542	0.4519
0.0691	8.0	960	3.6229	0.4583	0.4918	0.4583	0.4515
0.2158	9.0	1080	3.7731	0.4625	0.4741	0.4625	0.4595
0.1475	10.0	1200	4.1775	0.4667	0.4678	0.4667	0.4540
0.1593	11.0	1320	4.1295	0.45	0.4612	0.45	0.4453
0.0018	12.0	1440	4.2133	0.4417	0.4434	0.4417	0.4389
0.0006	13.0	1560	4.4056	0.4583	0.4736	0.4583	0.4567
0.0025	14.0	1680	4.6331	0.4667	0.4780	0.4667	0.4591
0.0001	15.0	1800	4.6897	0.4583	0.4582	0.4583	0.4530
0.0001	16.0	1920	4.6231	0.45	0.4549	0.45	0.4513
0.0003	17.0	2040	4.6506	0.4708	0.4706	0.4708	0.4695
0.0001	18.0	2160	4.6900	0.4708	0.4692	0.4708	0.4691
0.0001	19.0	2280	4.7095	0.4708	0.4695	0.4708	0.4657
0.0047	20.0	2400	4.7870	0.4625	0.4649	0.4625	0.4590
0.0001	21.0	2520	4.8568	0.4708	0.4697	0.4708	0.4687
0.0001	22.0	2640	4.9106	0.4708	0.4748	0.4708	0.4675
0.0001	23.0	2760	4.8998	0.4667	0.4699	0.4667	0.4627
0.0001	24.0	2880	4.9071	0.4708	0.4742	0.4708	0.4671

Framework versions

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