Codebase for Embedding Propagation: Smoother Manifold for Few-Shot Classification. This is a ServiceNow Research project that was started at Element AI.
In the pre-training process, is it possible to directly consider the three data sets to pre-train together? Shouldn't it be a training set of data that is pre-trained and then fine-tuned for testing on that data set?
In the pre-training process, is it possible to directly consider the three data sets to pre-train together? Shouldn't it be a training set of data that is pre-trained and then fine-tuned for testing on that data set?