487ecc79edb6b409fe9346206bda54a7

This model is a fine-tuned version of FacebookAI/xlm-roberta-large-finetuned-conll03-english on the contemmcm/cls_mmlu dataset. It achieves the following results on the evaluation set:

• Loss: 1.3869
• Data Size: 1.0
• Epoch Runtime: 72.9600
• 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	1.3901	0	2.7944	0.2447	0.2088
No log	1	438	1.4093	0.0078	3.7307	0.2527	0.1008
No log	2	876	1.4062	0.0156	4.9736	0.2533	0.1411
No log	3	1314	1.3899	0.0312	6.7619	0.2533	0.1011
No log	4	1752	1.3939	0.0625	9.7363	0.2487	0.0996
0.0784	5	2190	1.3959	0.125	14.2269	0.2527	0.1008
0.1856	6	2628	1.3895	0.25	22.8494	0.2487	0.0996
1.3987	7	3066	1.3936	0.5	40.3697	0.2487	0.0996
1.3973	8.0	3504	1.3898	1.0	73.9324	0.2487	0.0996
1.3989	9.0	3942	1.3930	1.0	73.9347	0.2527	0.1008
1.3956	10.0	4380	1.3890	1.0	73.6152	0.2527	0.1008
1.3951	11.0	4818	1.3931	1.0	72.8311	0.2533	0.1011
1.4003	12.0	5256	1.3914	1.0	72.6233	0.2527	0.1008
1.3948	13.0	5694	1.3875	1.0	72.3561	0.2533	0.1011
1.3987	14.0	6132	1.3866	1.0	72.1960	0.2527	0.1008
1.3945	15.0	6570	1.3908	1.0	71.9517	0.2527	0.1008
1.3957	16.0	7008	1.3850	1.0	72.9233	0.2533	0.1011
1.3911	17.0	7446	1.3880	1.0	72.6568	0.2527	0.1008
1.3946	18.0	7884	1.3913	1.0	72.0914	0.2527	0.1008
1.3944	19.0	8322	1.3862	1.0	71.4939	0.2533	0.1011
1.3935	20.0	8760	1.3869	1.0	72.9600	0.2487	0.0996

Framework versions

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