This is the Finnish GPT3 XL model (https://huggingface.co/TurkuNLP/gpt3-finnish-xl/) finetuned for simplification. Finetuning was done according to these instructions: https://github.com/spyysalo/instruction-finetune. The instruction for simplification is "Mukauta selkosuomeksi\n\n".
Paper: Towards Automatic Finnish Text Simplification (Dmitrieva & Tiedemann, DeTermIt-WS 2024).
The finetuning data can be obtained here: http://urn.fi/urn:nbn:fi:lb-2024011703. If you wish to replicate the results, you can find the training, validation, and testing sentence pairs' ids in the "splits.zip" archive in this repository. The ids contain the following information: "{regular text id}__{simple text id}__{sentence pair number}".
- Downloads last month
- 52
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.