2c0a19a4900712f9c5165af4e905d6ed

This model is a fine-tuned version of google-bert/bert-base-multilingual-cased on the nyu-mll/glue [mrpc] dataset. It achieves the following results on the evaluation set:

• Loss: 0.6666
• Data Size: 1.0
• Epoch Runtime: 7.7415
• Accuracy: 0.8243
• F1 Macro: 0.8012
• Rouge1: 0.8243
• Rouge2: 0.0
• Rougel: 0.8243
• Rougelsum: 0.8249

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	Rouge1	Rouge2	Rougel	Rougelsum
No log	0	0	0.6722	0	1.3019	0.6592	0.4163	0.6592	0.0	0.6592	0.6592
No log	1	114	0.6300	0.0078	2.1253	0.6645	0.4075	0.6651	0.0	0.6639	0.6645
No log	2	228	0.6625	0.0156	1.7367	0.6651	0.3994	0.6657	0.0	0.6645	0.6651
No log	3	342	0.7779	0.0312	2.1783	0.6651	0.3994	0.6657	0.0	0.6645	0.6651
0.0202	4	456	0.6098	0.0625	2.5420	0.6781	0.4612	0.6781	0.0	0.6775	0.6775
0.0202	5	570	0.6100	0.125	2.7685	0.6893	0.5242	0.6893	0.0	0.6887	0.6887
0.0202	6	684	0.5732	0.25	3.6647	0.7093	0.5623	0.7093	0.0	0.7093	0.7087
0.1435	7	798	0.3960	0.5	5.0147	0.8308	0.8133	0.8308	0.0	0.8308	0.8308
0.3789	8.0	912	0.3806	1.0	8.1707	0.8396	0.8138	0.8402	0.0	0.8396	0.8396
0.1992	9.0	1026	0.4424	1.0	8.0040	0.8096	0.7655	0.8096	0.0	0.8101	0.8096
0.1691	10.0	1140	0.6077	1.0	7.3394	0.8208	0.7813	0.8208	0.0	0.8208	0.8208
0.0876	11.0	1254	0.8537	1.0	7.3742	0.8107	0.7783	0.8101	0.0	0.8107	0.8107
0.0935	12.0	1368	0.6666	1.0	7.7415	0.8243	0.8012	0.8243	0.0	0.8243	0.8249

Framework versions

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