dd0b8c253651e49be7945051e101d6b0

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

• Loss: 2.1684
• Data Size: 1.0
• Epoch Runtime: 13.1481
• Accuracy: 0.2726
• F1 Macro: 0.2649

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	1.3914	0	0.9577	0.2527	0.1653
No log	1	438	1.4253	0.0078	1.3710	0.2527	0.1008
No log	2	876	1.3962	0.0156	1.3770	0.2447	0.1845
No log	3	1314	1.3964	0.0312	1.5808	0.2540	0.1194
No log	4	1752	1.3928	0.0625	1.9794	0.2460	0.1061
0.078	5	2190	1.3909	0.125	2.6827	0.2586	0.1691
0.1835	6	2628	1.3930	0.25	4.2590	0.2527	0.1385
1.3879	7	3066	1.3890	0.5	7.4244	0.2713	0.1906
1.3635	8.0	3504	1.3944	1.0	13.6031	0.2646	0.2333
1.2181	9.0	3942	1.4573	1.0	13.2843	0.2753	0.2569
0.877	10.0	4380	1.8970	1.0	13.5860	0.2786	0.2738
0.5663	11.0	4818	2.1684	1.0	13.1481	0.2726	0.2649

Framework versions

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