1441db9ea7cc185e2dede8ae156d4cbd

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

• Loss: 1.4751
• Data Size: 1.0
• Epoch Runtime: 28.8673
• Accuracy: 0.2420
• F1 Macro: 0.1040

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	1.4167	0	1.4470	0.25	0.2156
No log	1	438	1.4783	0.0078	1.8746	0.2566	0.1479
No log	2	876	1.3949	0.0156	1.7813	0.2540	0.1723
No log	3	1314	1.4841	0.0312	2.2258	0.2493	0.1025
No log	4	1752	1.3935	0.0625	3.1403	0.2527	0.1008
0.0787	5	2190	1.3953	0.125	4.7807	0.2527	0.1008
0.1885	6	2628	1.4284	0.25	8.2006	0.2487	0.0996
1.414	7	3066	1.3942	0.5	15.0427	0.2487	0.0996
1.4013	8.0	3504	1.4751	1.0	28.8673	0.2420	0.1040

Framework versions

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