This repository contains the code and pre-trained models for our EMNLP2022 Findings paper LEMON: Language-Based Environment Manipulation via Execution-guided Pre-training
The data is in the release. Please unzip it and put it in the lemon_data folder.
Run the following command to preprocess the data:
bash preprocess_pretrain.bat
Then run the following command to pre-train the model:
bash pretrain.sh
Run the following command to preprocess the data:
bash preprocess_finetune.bat
Then run the following command to fine-tune the model:
bash finetune.sh