bert_large_xsum_samsum2

This model is a fine-tuned version of alexdg19/bert_large_xsum_samsum on the samsum dataset. It achieves the following results on the evaluation set:

• Loss: 1.1949
• Rouge1: 0.6112
• Rouge2: 0.3855
• Rougel: 0.5301
• Rougelsum: 0.5296
• Gen Len: 30.5427

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: 2e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	41	0.9966	0.6323	0.416	0.5587	0.5598	26.9573
No log	2.0	82	1.0976	0.6279	0.413	0.5569	0.5583	27.8171
No log	3.0	123	1.1576	0.6236	0.4141	0.553	0.5537	29.5183
No log	4.0	164	1.1998	0.6148	0.3948	0.5402	0.541	30.5061
No log	5.0	205	1.1949	0.6112	0.3855	0.5301	0.5296	30.5427

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1