Rationale

[schopra6]

Still in improvement phase
readme file has instructions to run the flan t5 fine-tuned model

[nfelnlp]

Looks great! Could you upload the fine-tuned model to the DFKI cloud?
For the next steps, it's more important to use your existing one instead of training a new one. Then we can extend the rationalize operation by adding an option to use the FLAN-T5 instead of the GPT-Neo.