3a477aa447129a36ae5055b0c6b5cb4d

This model is a fine-tuned version of albert/albert-xlarge-v1 on the contemmcm/cls_20newsgroups dataset. It achieves the following results on the evaluation set:

• Loss: 3.0035
• Data Size: 1.0
• Epoch Runtime: 121.7274
• Accuracy: 0.0444
• F1 Macro: 0.0042

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	3.4098	0	8.6806	0.0444	0.0138
No log	1	499	3.1155	0.0078	9.8110	0.0507	0.0328
0.0327	2	998	3.0755	0.0156	10.6222	0.0509	0.0337
0.0569	3	1497	3.1373	0.0312	12.4048	0.0423	0.0041
0.1058	4	1996	3.0296	0.0625	16.1467	0.0517	0.0049
3.034	5	2495	3.0197	0.125	23.0827	0.0552	0.0068
3.0047	6	2994	3.0078	0.25	37.1461	0.0532	0.0051
3.0071	7	3493	3.0130	0.5	65.4711	0.0449	0.0043
3.0085	8.0	3992	2.9979	1.0	121.9411	0.0496	0.0047
3.0049	9.0	4491	3.0049	1.0	121.9029	0.0449	0.0043
3.0003	10.0	4990	3.0041	1.0	121.7119	0.0444	0.0042
3.0014	11.0	5489	2.9992	1.0	121.8089	0.0502	0.0048
2.9999	12.0	5988	3.0035	1.0	121.7274	0.0444	0.0042

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.1