42a65fd48de66f192cd402a6b27a571f

This model is a fine-tuned version of studio-ousia/mluke-large on the contemmcm/cls_20newsgroups dataset. It achieves the following results on the evaluation set:

• Loss: 0.4315
• Data Size: 1.0
• Epoch Runtime: 108.3309
• Accuracy: 0.9007
• F1 Macro: 0.8998

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	2.9973	0	7.4357	0.0524	0.0092
No log	1	499	3.0007	0.0078	8.5477	0.0572	0.0095
0.0301	2	998	2.9950	0.0156	10.0454	0.0832	0.0156
0.0542	3	1497	2.9034	0.0312	12.3663	0.2563	0.1907
0.0991	4	1996	1.8552	0.0625	16.4831	0.5010	0.4385
1.4812	5	2495	0.9915	0.125	23.1119	0.7122	0.7016
0.7185	6	2994	0.6801	0.25	35.9759	0.7870	0.7783
0.51	7	3493	0.4781	0.5	60.4961	0.8490	0.8476
0.3528	8.0	3992	0.4085	1.0	111.7089	0.8821	0.8826
0.3077	9.0	4491	0.4047	1.0	108.8379	0.8808	0.8786
0.2234	10.0	4990	0.4096	1.0	109.3024	0.8866	0.8850
0.1961	11.0	5489	0.4076	1.0	109.2466	0.8947	0.8940
0.1454	12.0	5988	0.4393	1.0	108.9655	0.8909	0.8904
0.1222	13.0	6487	0.4315	1.0	108.3309	0.9007	0.8998

Framework versions

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