You can deploy this model and start earning money today!

martinjosifoski/
SynthIE

Views

No views yet

This repository hosts the pre-trained models from the paper Exploiting Asymmetry for Synthetic Training Data Generation: SynthIE and the Case of Information Extraction. It is a companion to the project's homepage and GitHub repository, which contains all the details.

The repository contains 5 models:

SynthIE-base-FE (synthie_base_fe.ckpt): FLAN-T5-base, finetuned on the SynthIE-code dataset following the fully-expanded output linearization
SynthIE-base-SC (synthie_base_sc.ckpt): FLAN-T5-base, finetuned on the SynthIE-code dataset following the subject-collapsed linearization
SynthIE-large-FE (synthie_large_fe.ckpt): FLAN-T5-large, finetuned on the SynthIE-code dataset following the fully-expanded linearization
GenIE-base-FE (genie_base_fe.ckpt): FLAN-T5-base, finetuned on the REBEL dataset following the fully-expanded output linearization
GenIE-base-SC (genie_base_sc.ckpt): FLAN-T5-base, finetuned on the REBEL dataset following the subject-collapsed output linearization

The demo notebook in the project's GitHub repository provides instructions on how to download, load and use the models (as well as other resources released by the paper such as the datasets).
For more information, please refer to the project's GitHub repository and the paper.

martinjosifoski/SynthIE

martinjosifoski/
SynthIE