6b5671fa58dfcf87f1e0395d2fbaa316

This model is a fine-tuned version of facebook/opt-125m on the contemmcm/cls_20newsgroups dataset. It achieves the following results on the evaluation set:

• Loss: 0.5226
• Data Size: 1.0
• Epoch Runtime: 53.5998
• Accuracy: 0.8871
• F1 Macro: 0.8873

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	3.1454	0	4.6086	0.0575	0.0313
No log	1	499	3.0989	0.0078	5.0792	0.0701	0.0459
0.0313	2	998	2.8638	0.0156	5.3439	0.1346	0.0952
0.0529	3	1497	1.4615	0.0312	6.4227	0.5209	0.4721
0.0684	4	1996	0.7903	0.0625	7.9412	0.7205	0.7079
0.783	5	2495	0.6778	0.125	11.1595	0.7863	0.7890
0.5063	6	2994	0.5992	0.25	17.2904	0.8145	0.8114
0.3918	7	3493	0.4005	0.5	29.6062	0.8662	0.8669
0.2969	8.0	3992	0.3616	1.0	53.8036	0.8808	0.8808
0.2446	9.0	4491	0.3890	1.0	53.8079	0.8783	0.8771
0.136	10.0	4990	0.4394	1.0	53.7154	0.8823	0.8807
0.1189	11.0	5489	0.4702	1.0	54.3447	0.8833	0.8834
0.1124	12.0	5988	0.5226	1.0	53.5998	0.8871	0.8873

Framework versions

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