Execution instructions for Colbert: Refer to folder https://github.com/sameerp30/Structured-generation/tree/main/Colbert_files

SBert_top_3.py - Give the argument of num_negatives in the following format. Python3 <num_negatives>

Colbert_train_triple.py - Give the argument of num_negatives in the following format. Python3 <num_negatives>

prepare_index_data.py

prepare_train_data.py

Finally, the training file. train_colbert.py: Give path of indexing, query file.

Give the path of the Colbert-v2 checkpoint (https://downloads.cs.stanford.edu/nlp/data/colbert/colbertv2/colbertv2.0.tar.gz) for the argument checkpoint.

Data Folder All the scrapped and alternate flags data is in the folder https://github.com/sameerp30/Structured-generation/tree/main/flags_and_description_data. finals_index.tsv file in the same folder is a corpus of descriptions used for IR