bart-base-combined

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

• Loss: 2.2529
• Rouge1: 39.474
• Rouge2: 15.406
• Rougel: 27.6651
• Rougelsum: 36.2084
• Gen Len: 75.5717

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.5384	1.0	5435	2.3837	38.4347	14.2933	26.4507	35.1406	80.9404
2.2422	2.0	10870	2.3314	38.8455	14.6495	27.0452	35.4609	77.2153
2.0497	3.0	16305	2.3032	39.0212	14.8913	27.3328	35.7033	75.7681

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2