Fine Tuning Facebook-bart-large-cnn on the Samsum Text Summarization dataset

Training and benchmarking text-summarization models against Rouge score on the Samsum Dataset.

Results :

Improvement after Fine-Tuning

Rouge1 : 30.6 % Improvement
Rouge2 : 103 % Improvement
RougeL : 33.18 % Improvement
RougeLSum : 33.18 % Improvement

The fine-tuned model can be found at https://huggingface.co/dhivyeshrk/bart-large-cnn-samsum Runtime Logs and GPU utilization can be found in wandb_logs.pdf Trained on Nvidia Tesla P100

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 1

Streamlit App

Open Streamlit_App_Text_Summarizer.ipynb for a live demo It is recommended to use colab.