Open francescopisu opened 3 months ago
Hey @francescopisu! Thanks for opening an issue. :)
We do not intend to release the training pipeline in the near future, mostly because the training pipeline is tightly intertwined with internal Microsoft infrastructure. However, precisely how we trained the models is described in detail in the paper, where we tried to not leave out any details.
Thanks for this awesome work.
Any plans for the release of the training code?