97553293ba1bc6618ae6ebe824f8845d

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

• Loss: 0.4724
• Data Size: 1.0
• Epoch Runtime: 19.3074
• Accuracy: 0.8740
• F1 Macro: 0.8747

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	3.0012	0	1.7520	0.0539	0.0138
No log	1	499	2.9952	0.0078	2.0813	0.0585	0.0166
0.03	2	998	2.9521	0.0156	2.4837	0.1018	0.0515
0.0537	3	1497	2.5088	0.0312	2.7617	0.2797	0.2259
0.0908	4	1996	1.8004	0.0625	3.3996	0.4821	0.4249
1.5336	5	2495	1.2249	0.125	4.2795	0.6212	0.5768
0.8822	6	2994	0.7759	0.25	6.5032	0.7573	0.7405
0.6503	7	3493	0.5325	0.5	10.8104	0.8334	0.8320
0.3947	8.0	3992	0.4137	1.0	19.7697	0.8687	0.8679
0.3105	9.0	4491	0.4219	1.0	19.2888	0.8664	0.8651
0.1999	10.0	4990	0.4450	1.0	19.2375	0.8730	0.8725
0.1433	11.0	5489	0.4563	1.0	19.2985	0.8783	0.8784
0.1458	12.0	5988	0.4724	1.0	19.3074	0.8740	0.8747

Framework versions

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