7b4448c0bfa4fc54b37fbefb210ab573

This model is a fine-tuned version of meta-llama/Llama-3.2-1B on the contemmcm/cls_20newsgroups dataset. It achieves the following results on the evaluation set:

• Loss: 3.0380
• Data Size: 1.0
• Epoch Runtime: 261.7490
• Accuracy: 0.8521
• F1 Macro: 0.8528

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	19.7829	0	19.3171	0.0484	0.0209
No log	1	499	24.6663	0.0078	21.1424	0.0607	0.0138
0.2437	2	998	13.6939	0.0156	23.6445	0.0507	0.0048
0.2705	3	1497	12.1158	0.0312	28.6147	0.0421	0.0041
0.4183	4	1996	12.0830	0.0625	35.9813	0.0567	0.0091
12.0557	5	2495	11.8832	0.125	51.1226	0.0675	0.0126
11.1365	6	2994	9.4624	0.25	81.5520	0.1323	0.0525
5.2982	7	3493	4.0158	0.5	140.7765	0.6797	0.6677
2.1718	8.0	3992	2.1307	1.0	261.5188	0.8367	0.8280
1.1593	9.0	4491	1.9714	1.0	260.7197	0.8498	0.8519
0.9231	10.0	4990	2.1736	1.0	263.5476	0.8553	0.8539
0.5672	11.0	5489	2.9703	1.0	265.3278	0.8614	0.8614
0.4735	12.0	5988	2.9901	1.0	262.1528	0.8538	0.8547
0.5365	13.0	6487	3.0380	1.0	261.7490	0.8521	0.8528

Framework versions

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