I try to train the Transformers, and I get this errors: RuntimeError: view size is not compatible with input tensor's size and stride (at least one dimension spans across two contiguous subspaces). Use .reshape(...) instead.

dome272 / VQGAN-pytorch

Pytorch implementation of VQGAN (Taming Transformers for High-Resolution Image Synthesis) (https://arxiv.org/pdf/2012.09841.pdf)

MIT License

443 stars 73 forks source link

I try to train the Transformers, and I get this errors: RuntimeError: view size is not compatible with input tensor's size and stride (at least one dimension spans across two contiguous subspaces). Use .reshape(...) instead. #21

Closed F4k3r22 closed 2 months ago

F4k3r22 commented 3 months ago

I try to train the Transformers, and I get this errors: RuntimeError: view size is not compatible with input tensor's size and stride (at least one dimension spans across two contiguous subspaces). Use .reshape(...) instead. fuf