45d13b360b3b1dc775f76b31ea41d39e

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

• Loss: 1.3881
• Data Size: 1.0
• Epoch Runtime: 63.8954
• 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	1.4978	0	1.8739	0.2527	0.1360
No log	1	438	1.4446	0.0078	2.6062	0.2586	0.2206
No log	2	876	1.4745	0.0156	2.8690	0.2553	0.2338
No log	3	1314	1.3963	0.0312	3.9754	0.2580	0.1865
No log	4	1752	1.6268	0.0625	6.0347	0.2527	0.1008
0.0817	5	2190	1.3969	0.125	9.7687	0.2453	0.0985
0.1851	6	2628	1.4005	0.25	17.5922	0.2487	0.0996
1.3961	7	3066	1.3946	0.5	33.1709	0.2487	0.0996
1.3915	8.0	3504	1.3906	1.0	64.6564	0.2487	0.0996
1.387	9.0	3942	1.3966	1.0	64.4813	0.2527	0.1008
1.3874	10.0	4380	1.3908	1.0	64.1444	0.2527	0.1008
1.3875	11.0	4818	1.3894	1.0	64.3928	0.2533	0.1011
1.3878	12.0	5256	1.3879	1.0	64.5031	0.2527	0.1008
1.3892	13.0	5694	1.3897	1.0	63.6331	0.2533	0.1011
1.3892	14.0	6132	1.3898	1.0	64.2052	0.2527	0.1008
1.3882	15.0	6570	1.3897	1.0	64.2517	0.2527	0.1008
1.3885	16.0	7008	1.3881	1.0	63.8954	0.2527	0.1008

Framework versions

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