09cbd937b2a4ef22746fd7293d765391

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

• Loss: 1.1043
• Data Size: 1.0
• Epoch Runtime: 1137.3446
• Accuracy: 0.3182
• F1 Macro: 0.1609
• Rouge1: 0.3184
• Rouge2: 0.0
• Rougel: 0.3182
• Rougelsum: 0.3183

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	1.2907	0	7.9126	0.3540	0.1751	0.3539	0.0	0.3540	0.3538
1.1102	1	12271	0.9334	0.0078	19.0122	0.5456	0.5088	0.5456	0.0	0.5459	0.5455
0.7655	2	24542	0.6558	0.0156	26.9505	0.7339	0.7338	0.7339	0.0	0.7339	0.7338
0.6731	3	36813	0.6467	0.0312	44.3297	0.75	0.7465	0.7497	0.0	0.7502	0.75
0.6613	4	49084	0.5680	0.0625	80.0463	0.7793	0.7773	0.7793	0.0	0.7794	0.7795
0.5785	5	61355	0.5429	0.125	149.8707	0.7830	0.7792	0.7829	0.0	0.7830	0.7830
0.7309	6	73626	0.7173	0.25	289.3004	0.6990	0.6976	0.6989	0.0	0.6989	0.6989
0.6779	7	85897	0.7155	0.5	581.7114	0.7050	0.7058	0.7048	0.0	0.7048	0.7050
1.1055	8.0	98168	1.0986	1.0	1133.6875	0.3545	0.1745	0.3544	0.0	0.3545	0.3543
1.1071	9.0	110439	1.1043	1.0	1137.3446	0.3182	0.1609	0.3184	0.0	0.3182	0.3183

Framework versions

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