a9556ccd2d0850daf40fa061fcda6fd3

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

• Loss: 1.3881
• Data Size: 1.0
• Epoch Runtime: 22.9238
• Accuracy: 0.2533
• F1 Macro: 0.1011

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	1.4292	0	1.2894	0.2374	0.1563
No log	1	438	1.4016	0.0078	1.9518	0.2527	0.1008
No log	2	876	1.4126	0.0156	1.8195	0.2440	0.1130
No log	3	1314	1.4259	0.0312	2.2299	0.2533	0.1011
No log	4	1752	1.3928	0.0625	2.9440	0.2487	0.0996
0.0788	5	2190	1.3935	0.125	4.1509	0.2553	0.1262
0.1857	6	2628	1.3927	0.25	6.9383	0.2487	0.0996
1.392	7	3066	1.3910	0.5	12.4288	0.2487	0.0996
1.4005	8.0	3504	1.3903	1.0	24.4327	0.2487	0.0996
1.3966	9.0	3942	1.3912	1.0	23.3583	0.2527	0.1008
1.3958	10.0	4380	1.3920	1.0	23.0242	0.2453	0.0985
1.3999	11.0	4818	1.3911	1.0	23.5226	0.2533	0.1011
1.4008	12.0	5256	1.3900	1.0	24.1910	0.2527	0.1008
1.3969	13.0	5694	1.3878	1.0	23.5728	0.2533	0.1011
1.4003	14.0	6132	1.3897	1.0	22.9778	0.2527	0.1008
1.3996	15.0	6570	1.3905	1.0	24.2439	0.2527	0.1008
1.3978	16.0	7008	1.3880	1.0	23.2027	0.2527	0.1008
1.3921	17.0	7446	1.3869	1.0	23.6137	0.2453	0.0985
1.3943	18.0	7884	1.3900	1.0	24.3504	0.2527	0.1008
1.3965	19.0	8322	1.3880	1.0	23.3937	0.2533	0.1011
1.3999	20.0	8760	1.3876	1.0	23.2422	0.2487	0.0996
1.4054	21.0	9198	1.3851	1.0	23.8980	0.2487	0.0996
1.396	22.0	9636	1.3887	1.0	22.7160	0.2533	0.1011
1.3983	23.0	10074	1.3894	1.0	23.1777	0.2533	0.1011
1.3971	24.0	10512	1.3910	1.0	24.0075	0.2487	0.0996
1.3964	25.0	10950	1.3881	1.0	22.9238	0.2533	0.1011

Framework versions

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