bert_large_xsum_samsum

This model is a fine-tuned version of facebook/bart-large-xsum on the samsum dataset. It achieves the following results on the evaluation set:

• Loss: 1.9030
• Rouge1: 0.5083
• Rouge2: 0.2528
• Rougel: 0.41
• Rougelsum: 0.4105
• Gen Len: 29.0183

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	1.6008	0.4779	0.2349	0.4058	0.4056	21.1037
No log	2.0	82	1.5804	0.5104	0.2526	0.4242	0.4239	24.689
No log	3.0	123	1.7310	0.5148	0.253	0.4162	0.4155	28.0793
No log	4.0	164	1.7974	0.5019	0.2443	0.4127	0.4125	25.189
No log	5.0	205	1.9030	0.5083	0.2528	0.41	0.4105	29.0183

Framework versions

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