SumTO @ FNS 2020

The repository include the evaluation code fot the SumTO summarization system proposed for the FNS 2020 Shared Task.

Evaluation script

• Summarizer.py include the code for the Summarizer python object. It is able to initialize the model and perform the summarization using the .summarize() function
• summarize.py contains the code to initialize and apply the model to pre-parsed input data collections.
• In summarize.py: DATA_DIR and TEST_DIR should be set according to your environment configuration
• In summarize.py: YourSystemID should be set according to your output folder (it will contain the summarized documents at the end of the summarization process)
• components/Dataset.py contains the Dataset class exploited by the summarization algorithm to predict the summaries.
• create_dataset.py contains the instructions to create and store the Dataset object (this version is intended explictly for the test set).

Pre-trained Financial model

Available at https://huggingface.co/morenolq/SumTO_FNS2020 or using the transformers python library with the tag morenolq/SumTO_FNS2020

Citation (Coming Soon)

La Quatra, M., & Cagliero, L. (2020, December). End-to-end Training For Financial Report Summarization. In Proceedings of the 1st Joint Workshop on Financial Narrative Processing and MultiLing Financial Summarisation (pp. 118-123).

https://www.aclweb.org/anthology/2020.fnp-1.20.pdf