bart-large-oposum2

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

• Loss: 3.2051
• Rouge1: 0.6185
• Rouge2: 0.6021
• Rougel: 0.6157
• Rougelsum: 0.6157
• Gen Len: 18.932

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: 16
• total_train_batch_size: 64
• 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
No log	0.9984	156	3.5332	0.5855	0.5682	0.5823	0.5824	22.576
No log	1.9968	312	3.2461	0.6109	0.5946	0.6084	0.6083	19.284
No log	2.9952	468	3.2051	0.6185	0.6021	0.6157	0.6157	18.932

Framework versions

• PEFT 0.10.1.dev0
• Transformers 4.40.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1