f5dab6d526d3c97f29b9bbed85419a35

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

• Loss: 0.7889
• Data Size: 1.0
• Epoch Runtime: 7.5318
• Accuracy: 0.8154
• F1 Macro: 0.7856
• Rouge1: 0.8154
• Rouge2: 0.0
• Rougel: 0.8160
• Rougelsum: 0.8160

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	0.6655	0	1.3459	0.6616	0.4096	0.6621	0.0	0.6616	0.6616
No log	1	114	0.6370	0.0078	1.8298	0.6651	0.3994	0.6657	0.0	0.6645	0.6651
No log	2	228	0.6580	0.0156	1.6635	0.6651	0.3994	0.6657	0.0	0.6645	0.6651
No log	3	342	0.6526	0.0312	2.0175	0.6651	0.3994	0.6657	0.0	0.6645	0.6651
0.0206	4	456	0.6426	0.0625	2.2644	0.6651	0.3994	0.6657	0.0	0.6645	0.6651
0.0206	5	570	0.6098	0.125	2.5314	0.6816	0.4547	0.6816	0.0	0.6816	0.6816
0.0206	6	684	0.4724	0.25	3.3047	0.7665	0.7560	0.7665	0.0	0.7665	0.7665
0.1303	7	798	0.4271	0.5	4.6421	0.7966	0.7770	0.7966	0.0	0.7966	0.7972
0.3709	8.0	912	0.4185	1.0	7.9948	0.8343	0.8116	0.8343	0.0	0.8337	0.8349
0.2119	9.0	1026	0.4482	1.0	7.8722	0.8054	0.7847	0.8054	0.0	0.8060	0.8060
0.1643	10.0	1140	0.5311	1.0	7.3031	0.8107	0.7819	0.8107	0.0	0.8107	0.8107
0.1098	11.0	1254	0.7311	1.0	7.4085	0.8208	0.7953	0.8213	0.0	0.8213	0.8208
0.0965	12.0	1368	0.7889	1.0	7.5318	0.8154	0.7856	0.8154	0.0	0.8160	0.8160

Framework versions

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