bart-cnn-science-v3-e6

This model is a fine-tuned version of theojolliffe/bart-cnn-science on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.8057
• Rouge1: 53.7462
• Rouge2: 34.9622
• Rougel: 37.5676
• Rougelsum: 51.0619
• Gen Len: 142.0

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: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 6
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	398	0.9961	52.632	32.8104	35.0789	50.3747	142.0
1.174	2.0	796	0.8565	52.8308	32.7064	34.6605	50.3348	142.0
0.7073	3.0	1194	0.8322	52.2418	32.8677	36.1806	49.6297	141.5556
0.4867	4.0	1592	0.8137	53.5537	34.5404	36.7194	50.8394	142.0
0.4867	5.0	1990	0.7996	53.4959	35.1017	37.5143	50.9972	141.8704
0.3529	6.0	2388	0.8057	53.7462	34.9622	37.5676	51.0619	142.0

Framework versions

• Transformers 4.19.2
• Pytorch 1.11.0+cu113
• Datasets 2.2.2
• Tokenizers 0.12.1