delete unused parameters in config.json file

Update config.json

Corrected name on Readme file

Update base_model_name_or_path to point to the correct repository

adding the tokenizer to the repo

adding lighter version of GPT2-Einstein-EPFL using quantization 4bit

initial commit