5a2a5c0a0eb450885cd5fb1af9824857

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

• Loss: 1.3877
• Data Size: 1.0
• Epoch Runtime: 120.0157
• Accuracy: 0.2487
• F1 Macro: 0.0996

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	1.7784	0	3.2331	0.2460	0.1725
No log	1	438	1.4876	0.0078	4.3292	0.2407	0.2224
No log	2	876	1.4012	0.0156	5.1143	0.2427	0.2079
No log	3	1314	1.4484	0.0312	7.1618	0.2620	0.1681
No log	4	1752	1.3940	0.0625	11.0504	0.2527	0.1008
0.0825	5	2190	1.3985	0.125	18.0212	0.2453	0.0985
0.1948	6	2628	1.4229	0.25	32.7564	0.2487	0.0996
1.4657	7	3066	1.4154	0.5	61.8446	0.2453	0.0985
1.3901	8.0	3504	1.3900	1.0	121.1050	0.2487	0.0996
1.3862	9.0	3942	1.3896	1.0	120.6294	0.2527	0.1008
1.3877	10.0	4380	1.3903	1.0	119.7003	0.2527	0.1008
1.3873	11.0	4818	1.3876	1.0	120.3298	0.2533	0.1011
1.3881	12.0	5256	1.3880	1.0	120.7355	0.2527	0.1008
1.3885	13.0	5694	1.3871	1.0	119.5208	0.2487	0.0996
1.3885	14.0	6132	1.3877	1.0	120.3651	0.2527	0.1008
1.3874	15.0	6570	1.3895	1.0	120.5198	0.2527	0.1008
1.3875	16.0	7008	1.3881	1.0	119.8156	0.2527	0.1008
1.3837	17.0	7446	1.3867	1.0	120.1265	0.2487	0.0996
1.389	18.0	7884	1.3861	1.0	120.4231	0.2527	0.1008
1.3861	19.0	8322	1.3865	1.0	120.5597	0.2533	0.1011
1.3883	20.0	8760	1.3844	1.0	120.1635	0.2527	0.1008
1.387	21.0	9198	1.3880	1.0	120.1970	0.2527	0.1008
1.3874	22.0	9636	1.3847	1.0	120.1456	0.2533	0.1011
1.3855	23.0	10074	1.3882	1.0	119.9105	0.2533	0.1011
1.3856	24.0	10512	1.3877	1.0	120.0157	0.2487	0.0996

Framework versions

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