238d420795ba20c91f8df7017f199252

This model is a fine-tuned version of google-bert/bert-base-cased on the contemmcm/cls_mmlu dataset. It achieves the following results on the evaluation set:

• Loss: 1.3955
• Data Size: 1.0
• Epoch Runtime: 23.9501
• Accuracy: 0.2527
• F1 Macro: 0.1008

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	1.4089	0	1.2247	0.2460	0.1862
No log	1	438	1.4131	0.0078	1.6681	0.2533	0.1264
No log	2	876	1.4109	0.0156	1.7319	0.2453	0.1001
No log	3	1314	1.4117	0.0312	2.1588	0.2533	0.1011
No log	4	1752	1.3910	0.0625	3.4299	0.2487	0.1211
0.0782	5	2190	1.4038	0.125	4.4366	0.2613	0.1745
0.1853	6	2628	1.3971	0.25	7.0881	0.2487	0.0996
1.3963	7	3066	1.3947	0.5	12.3448	0.2487	0.0996
1.3967	8.0	3504	1.3955	1.0	23.9501	0.2527	0.1008

Framework versions

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