26231ce1d6fce53dbcb025e5ecabfe23

This model is a fine-tuned version of albert/albert-xxlarge-v2 on the contemmcm/cls_20newsgroups dataset. It achieves the following results on the evaluation set:

• Loss: 3.0017
• Data Size: 1.0
• Epoch Runtime: 229.5108
• Accuracy: 0.0491
• F1 Macro: 0.0047

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	3.7882	0	17.9369	0.0565	0.0137
No log	1	499	3.1992	0.0078	19.7926	0.0680	0.0357
0.0351	2	998	3.0048	0.0156	21.5011	0.0479	0.0090
0.0546	3	1497	3.0042	0.0312	25.1468	0.0433	0.0063
0.1028	4	1996	3.0112	0.0625	31.7864	0.0517	0.0049
3.0091	5	2495	2.9996	0.125	44.9239	0.0517	0.0049
2.9993	6	2994	3.0049	0.25	71.1133	0.0423	0.0041
2.9992	7	3493	3.0074	0.5	124.0731	0.0449	0.0043
3.001	8.0	3992	2.9960	1.0	229.8546	0.0517	0.0049
3.0025	9.0	4491	3.0016	1.0	229.4948	0.0491	0.0047
2.9993	10.0	4990	3.0023	1.0	229.5445	0.0444	0.0042
3.0001	11.0	5489	2.9989	1.0	229.4635	0.0502	0.0048
2.9986	12.0	5988	3.0017	1.0	229.5108	0.0491	0.0047

Framework versions

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