Pretraining On Subset of Dataset

[anantr37]

• pretrain_on_data.py runs pretraining using reconstruction on the datasets (defaults to pretraining on five datasets)
• process_dataset_files.py processes the .bin data files into csv files using the meta.json files (the .bin and meta.json files are in the dataset folder. I did not add these files as the large file size was causing issues)