6300d99089e5b971f7e554415b186af2

This model is a fine-tuned version of FacebookAI/xlm-roberta-large-finetuned-conll03-english on the contemmcm/cls_20newsgroups dataset. It achieves the following results on the evaluation set:

• Loss: 0.5352
• Data Size: 1.0
• Epoch Runtime: 88.3933
• Accuracy: 0.8798
• F1 Macro: 0.8786

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	3.0281	0	6.0063	0.0499	0.0146
No log	1	499	3.0534	0.0078	6.9088	0.0655	0.0153
0.0307	2	998	3.0019	0.0156	7.8730	0.0512	0.0049
0.0552	3	1497	2.5361	0.0312	9.8173	0.1497	0.0866
0.0941	4	1996	1.9715	0.0625	13.2614	0.3362	0.2463
1.655	5	2495	1.2215	0.125	18.9385	0.5509	0.4978
0.9772	6	2994	0.9406	0.25	30.1243	0.6752	0.6398
0.7882	7	3493	0.6904	0.5	50.2442	0.7913	0.7859
0.5434	8.0	3992	0.5522	1.0	89.4544	0.8382	0.8363
0.5133	9.0	4491	0.4544	1.0	88.4607	0.8558	0.8549
0.3696	10.0	4990	0.4541	1.0	88.8014	0.8667	0.8644
0.3615	11.0	5489	0.4739	1.0	88.4144	0.8667	0.8647
0.2583	12.0	5988	0.4722	1.0	89.1865	0.8737	0.8738
0.2758	13.0	6487	0.5547	1.0	89.1034	0.8579	0.8587
0.248	14.0	6986	0.5352	1.0	88.3933	0.8798	0.8786

Framework versions

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