66e309a800b59fb9eadb037e3314225a

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

• Loss: 1.4261
• Data Size: 1.0
• Epoch Runtime: 92.6445
• Accuracy: 0.2533
• F1 Macro: 0.1011

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	1.4406	0	2.6534	0.2560	0.2073
No log	1	438	1.4147	0.0078	3.4961	0.2453	0.0985
No log	2	876	1.3967	0.0156	4.0684	0.2553	0.1700
No log	3	1314	1.4272	0.0312	5.6066	0.2434	0.1542
No log	4	1752	1.3947	0.0625	8.6727	0.2527	0.1008
0.0788	5	2190	1.4027	0.125	14.0348	0.2527	0.1008
0.1858	6	2628	1.4047	0.25	25.3556	0.2487	0.0996
1.3988	7	3066	1.3920	0.5	48.2040	0.2487	0.0996
1.3937	8.0	3504	1.3949	1.0	93.4120	0.2527	0.1008
1.3931	9.0	3942	1.4019	1.0	93.0487	0.2533	0.1011
1.4114	10.0	4380	1.4016	1.0	92.3479	0.2527	0.1008
1.4259	11.0	4818	1.4261	1.0	92.6445	0.2533	0.1011

Framework versions

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