bart-finetuned-cnn-3

This model is a fine-tuned version of sshleifer/distilbart-xsum-12-3 on the cnn_dailymail dataset. It achieves the following results on the evaluation set:

• Loss: 2.0751
• Rouge1: 40.201
• Rouge2: 18.8482
• Rougel: 29.4439
• Rougelsum: 37.416
• Gen Len: 56.7545

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.276	1.0	8883	2.1762	39.6581	18.3333	28.7765	36.7688	58.5386
2.0806	2.0	17766	2.0909	40.0328	18.8026	29.417	37.3508	56.6804
1.9615	3.0	26649	2.0751	40.201	18.8482	29.4439	37.416	56.7545

Framework versions

• Transformers 4.16.2
• Pytorch 1.10.2+cu102
• Datasets 1.18.3
• Tokenizers 0.11.0