Show HN: BART-based summarizer trained on years of newsletter summaries

skriber.io

6 pointscopypirate7 comments

Summarizer is at skriber.io

Powered by BART and trained on my own dataset of a few thousand summaries. Let me know how it works for you :)

login: test

password: test123

Tried copy/paste but still got: Please enter a correct username and password. Note that both fields may be case-sensitive.

Oops
login: test pass: test123

There's an asterisk at the end of 123

It seems to work really well are you sharing the model and code?

Thanks, I definitely might once I tidy things up... though, there isn't much to share. The model is pre-trained on the CNN/Dailymail dataset and hosted on AWS SageMaker (I used the BART tutorial as a reference). Fine-tuned on my own article summaries using the transformers library. The backend is Django/Bootstrap. Really, nothing special except for the fine-tune dataset which took ages to clean.

It works well which is special. Open sourcing it good whether it’s clean or not

Reply on news.ycombinator.com