What is the difference between T5 and T0 models? I had read that T0 is T5 + LM. But as I know T5 uses encoder-decoder model like BART but BART can be used as LM so that's mean that T5 has a LM function too. So what is the difference between T5 and T0 then?
T0 starts with T5+LM as the base model but then performs multi-task finetuning on it using BIG-Bench tasks. As detailed in the Results section (Section 6) of their paper, T0 outperforms T5+LM on a number of tasks.