A small entirely-human-made dataset (of admittedly dubious quality) totaling 15k entries in size. I want to try and implement a "Guess the Instruction" task inspired by this paper to see how that affects model performance, even if it is relatively small compared to the other data we're feeding into Pyg.
A small entirely-human-made dataset (of admittedly dubious quality) totaling 15k entries in size. I want to try and implement a "Guess the Instruction" task inspired by this paper to see how that affects model performance, even if it is relatively small compared to the other data we're feeding into Pyg.