bart-base-samsum

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: 1.5834
• Rouge1: 47.552
• Rouge2: 24.8542
• Rougel: 40.56
• Rougelsum: 44.3423
• Gen Len: 17.8337

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: 8e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• 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
1.8884	1.0	1841	1.6325	47.0402	24.1542	39.6427	43.8472	18.5941
1.5081	2.0	3683	1.5834	47.552	24.8542	40.56	44.3423	17.8337
1.2216	3.0	5524	1.5855	48.2058	25.1623	40.9023	44.2822	17.901
1.0074	4.0	7366	1.6049	48.3145	25.2348	40.8688	44.4735	18.4829
0.8544	5.0	9205	1.6455	48.5926	25.3142	40.9144	44.6577	18.3924

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2