Skip to content

Latest commit

 

History

History
4 lines (3 loc) · 591 Bytes

File metadata and controls

4 lines (3 loc) · 591 Bytes

benchmarking-and-architectural-analysis-of-state-of-the-art-transformer-models

The code of my master thesis: Benchmarking and architectural analysis of state-of-the-art transformer models for book review summarization

The Code inside sum_models/finetuning is from the huggingface repository: https://github.com/huggingface/transformers and adjusted to finetune the Bart-Large-cnn model. Inside sum_models/experiments the jupyter notebooks can be found, showing the progess of the experiments. sum_models/experiments/revPrep contains the code, that is necessary for the notebooks to work.