d8e8d6f7d48903d983e6480cbb5c5585

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

• Loss: 0.7239
• Data Size: 1.0
• Epoch Runtime: 571.3262
• Accuracy: 0.7850
• F1 Macro: 0.7855
• Rouge1: 0.7851
• Rouge2: 0.0
• Rougel: 0.7849
• Rougelsum: 0.7853

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	1.1171	0	4.6729	0.3207	0.1868	0.3210	0.0	0.3206	0.3207
1.0954	1	12271	0.8933	0.0078	10.4538	0.5920	0.5887	0.5919	0.0	0.5918	0.5922
0.8373	2	24542	0.7553	0.0156	13.9618	0.6777	0.6697	0.6778	0.0	0.6780	0.6778
0.7039	3	36813	0.6737	0.0312	22.1968	0.7160	0.7113	0.7159	0.0	0.7161	0.7160
0.6721	4	49084	0.6217	0.0625	40.0914	0.7458	0.7450	0.7457	0.0	0.7462	0.7461
0.5656	5	61355	0.5808	0.125	75.6268	0.7599	0.7596	0.7598	0.0	0.7596	0.7601
0.5715	6	73626	0.5689	0.25	145.0585	0.7687	0.7692	0.7686	0.0	0.7688	0.7688
0.4756	7	85897	0.5482	0.5	284.8263	0.7868	0.7866	0.7866	0.0	0.7869	0.7868
0.4692	8.0	98168	0.5460	1.0	568.6943	0.7929	0.7929	0.7928	0.0	0.7930	0.7930
0.3914	9.0	110439	0.5184	1.0	566.2477	0.8032	0.8026	0.8032	0.0	0.8032	0.8032
0.3494	10.0	122710	0.5837	1.0	566.2320	0.7935	0.7933	0.7934	0.0	0.7937	0.7935
0.3036	11.0	134981	0.6292	1.0	572.6695	0.7966	0.7967	0.7965	0.0	0.7967	0.7969
0.2572	12.0	147252	0.6321	1.0	577.5898	0.7975	0.7971	0.7975	0.0	0.7975	0.7977
0.2571	13.0	159523	0.7239	1.0	571.3262	0.7850	0.7855	0.7851	0.0	0.7849	0.7853

Framework versions

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