TensorFlow and PyTorch implementation of "Meta-Transfer Learning for Few-Shot Learning" (CVPR2019)
TensorFlow and PyTorch implementation of "Meta-Transfer Learning for Few-Shot Learning" (CVPR2019)
the implementation for the hard-meta batch
Formula（4） for updating SS parameters
the code for choosing hard class-m？
Where is the code for choosing hard class-m ?
how to fix feature encoder weights during SS process
About the correspondence between the phases proposed in the paper and the phases in the code
About SS operation.
How do you update the base_leaner's parameters
“MAX_ITER=20000” during meta-test in Tensorflow code
in conv2d_mtl.py, the mtl_weight has 4-dimensional and the weight has 3-dimensional, new_mtl_weight = self.mtl_weight.expand(self.weight.shape) has error,is there some problem?