-
There is a _bug_ when transforming the **contravariant indices** of a tensor to **covariant indices** _for tensors of rank **higher than 2**_.
It seems that when transforming the tensor, the resultan…
-
### Issue Description
Many normalizing flows require the use of transpositions during the flow, allowing different transformations to operate on different dimensions. Since transpositions don't chang…
-
I have a question regarding the documentation of the implementation of the symmetry operations. This does not really belong here but I cannot think of another place to ask.
The docs have the image …
-
# Prerequisites
Please answer the following questions for yourself before submitting an issue.
- [x] I am running the latest code. Development is very rapid so there are no tagged versions as of…
nih23 updated
4 months ago
-
transpose causes Apply1 to be 20 times slower. Even though, almost all Apply1 ops are invariant to transpose. Exceptions include eg `tril`, but are rare. So, with following code:
```
local time = …
-
Hello, after looking at the code in patch merging part, we found the complex operation that slice the feature and concatenate them then go through the linear layer to reduce dimension from 4C to 2C is…
-
Hi @nicolasvasilache , all,
Before I adventure in writing a pass, I was wondering if you guys already thought about how transposition is handled in the code.
### How (I think) transposition is ha…
-
aten reference: [permute](https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/native/native_functions.yaml#L2230).
Semantics here are limited if we're to maintain the view functionality of…
-
The source codes are in https://pastebin.com/Fpn1Rqp8
Running:
Assertion `supported && "WGMMA type or shape is not supported"' failed.
The second operand is column major, but there is no way to…
-
Let's discuss and decide on some terminology to use consistently **for dense tensors**:
1) A tensor which is addressed by *n* indices is:
- [ ] a: *n*-dimensional
- [ ] b: *n*-mode
-…