Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Implement BatchLazyTensor

See original GitHub issue

There is a BatchRepeatLazyTensor, but no BatchLazyTensor that would allow batching of a number of arbitrary LazyTensors of the same size.

I’d like to use this to set up a BlockDiagLazyTensor with a base_lazy_tensor that is a batch of LazyTensors.

Issue Analytics

State:
Created 5 years ago
Comments:5 (1 by maintainers)

Top GitHub Comments

1reaction

gpleisscommented, Jan 25, 2019

I’m working on a unsqueezing option right now as part of the batch mode PR. I’ll try to get that component up soon. @Balandat

0reactions

Balandatcommented, Feb 22, 2019

Seems like we don’t need after all this as we can use CatLazyTensor on unsqueezed Lazies.

Top Results From Across the Web

LazyTensor — KeOps

LazyTensor encode numerical arrays through the combination of a symbolic, mathematical formula and a list of small data arrays. They can be used ......

Make DiagLazyTensor work with any batch size #361 - GitHub

I'm in agreement with this change for sure, a DiagLazyTensor created with a 2 x 3 x 4 tensor should represent a 2...

Source code for gpytorch.lazy.lazy_tensor

This probably won't have to be overwritten by LazyTensors, unless they use batch dimensions in a special way (e.g. BlockDiagLazyTensor, SumBatchLazyTensor) ...

Advanced Mini-Batching — pytorch_geometric documentation

All other tensors will just get concatenated in the first dimension without any further increasement of their values. However, there are a few...

Lazy Tensor Core - PyTorch Dev Discussions

Lazy Tensors in PyTorch is an active area of exploration, and this is a call for community involvement to discuss the requirements, implementation,...