carriex / recomp

RECOMP: Improving Retrieval-Augmented LMs with Compression and Selective Augmentation.
MIT License
99 stars 5 forks source link

Inference code #7

Closed shihabrashid-ucr closed 1 month ago

shihabrashid-ucr commented 8 months ago

Hi, thank you for the great work and for releasing the model cards! I was wondering is there any simple inference code with your model? From what I understand, in order to run the abstraction compressor with custom sets of documents for one query, I need to run the train_hf_summarization_model.py with the input file formatted like the one you provided. Is it possible to provide a simple function that will take k documents, the query and just output the abstractive summary given your trained model? If it is not possible to provide it, any guides to writing such function will be incredibly helpful!

carriex commented 4 months ago

Hi there! the code here provided is for inference.