8d03156aeec8fec2e351703176fdf038

This model is a fine-tuned version of FacebookAI/xlm-roberta-base on the contemmcm/cls_20newsgroups dataset. It achieves the following results on the evaluation set:

• Loss: 0.5209
• Data Size: 1.0
• Epoch Runtime: 45.8554
• Accuracy: 0.8788
• F1 Macro: 0.8765

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
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Accuracy	F1 Macro
No log	0	0	3.0081	0	3.6088	0.0476	0.0045
No log	1	499	3.0166	0.0078	4.3232	0.0993	0.0405
0.0298	2	998	3.0034	0.0156	4.4949	0.0519	0.0074
0.0542	3	1497	2.8232	0.0312	5.6376	0.1772	0.1023
0.0996	4	1996	1.6939	0.0625	7.0045	0.4458	0.3732
1.6155	5	2495	1.1898	0.125	9.8762	0.6137	0.5609
0.9197	6	2994	0.8469	0.25	15.3850	0.7414	0.7265
0.719	7	3493	0.6029	0.5	26.0814	0.8105	0.8094
0.4717	8.0	3992	0.4454	1.0	47.2016	0.8629	0.8606
0.3567	9.0	4491	0.4378	1.0	47.7746	0.8700	0.8674
0.2966	10.0	4990	0.4349	1.0	45.8373	0.8763	0.8741
0.2303	11.0	5489	0.4192	1.0	45.3867	0.8858	0.8849
0.1768	12.0	5988	0.4439	1.0	46.4635	0.8858	0.8861
0.1733	13.0	6487	0.4284	1.0	45.5862	0.8952	0.8946
0.1618	14.0	6986	0.5499	1.0	45.8690	0.8778	0.8768
0.1594	15.0	7485	0.5209	1.0	45.8554	0.8788	0.8765

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1