-
When you apply tf.gather to a random variable, edward seems to have trouble treating the result as a random variable. Here's an example: if `W` is a random variable of shape (100), then `tf.gather(W,n…
-
## What
Let's introduce dredd test for TensorShape.
## Why
To test shape update of circle models.
Related : https://github.com/Samsung/ONE/issues/10980
-
### 🐛 Describe the bug
I'm using Dynamo export on the [FullyConnectedNet](https://github.com/Project-MONAI/MONAI/blob/dev/monai/networks/nets/fullyconnectednet.py#L31) model by MONAI.
cc @ezyang…
-
One of the quite compelling new additions to the torch ecosystem is NestedTensors, see https://pytorch.org/docs/stable/nested.html.
Basically, this is a new primitive in the Torch ecosystem which a…
-
Transpose operations that earlier was applied for all Matmul/Conv operations was fused with it ops and covered with BLAS implementation. Maybe it's possible to convert in the similar way other common …
-
### Use Case
Most solid material models require tensor operations to compute stress $\mathbf{S}$ and tangent matrix $\mathrm{C}$. The tensors are usually 2nd- and 4th-order fixed-size, `nsd x nsd` an…
-
### 🐛 Describe the bug
The following script attempts to fuse two custom operations together into a single custom op. One of the original ops, plus the fused op have multiple outputs. The resultin…
-
While experimenting with PyTorch's DTensor functionality in my single node "cpu" environment, I've encountered an unexpected limitation. Specifically, in the provided code example, I'm attempting to u…
-
The `Promoted` type can cause a mismatch with the runtime promotion in certain cases when combining tensors with scalar values.
The reason is, that the type promotion , somewhat unintuitively, does…
-
Hi Choy. Thanks for your great work on this project! I'm trying to calculate the second derivatives of sparse tensor operations (e.g. sparse convolution). I found that the backward operation of sparse…