-
Minified from internal, after patching in https://github.com/pytorch/pytorch/pull/132524:
```
@torch._dynamo.config.patch(
capture_dynamic_output_shape_ops=True, capture_scalar_outputs=…
-
# Feature request:
I think it would be quite useful to have a function that applies a coordinate transformation to a tensor component, or to all components. Some specialized functions were implemen…
-
Out of the 8 cases ({+, -, *, /} × { LHS, RHS }), the following 2 cases does not work:
- `scalar * tensor` fails to compile
- `scalar / tensor` gives incorrect results by computi…
-
If I write x[i]*x[i], then by the documented rules of Indexed expressions it seems there is supposed to be an implied sum, since the index i is present twice. However this gets automatically simplifie…
-
https://github.com/dealii/dealii/pull/8033#discussion_r283159683 refers
Plurals of class names don't render. The following changes have been proposed:
`Expressions`->`Expression objects`
`Tensors…
-
Hi,
I tried to build with ROCm support and it's fail.
Build output:
```
quentin@quentin-pc:~/Documents/stable-diffusion.cpp/build$ cmake .. -G "Ninja" -DCMAKE_C_COMPILER=$(hipconfig -l)/clang …
-
Why does this repository appear empty?
I have some questions regarding the paper titled "Optimizing Deep Learning Inference via Global Analysis and Tensor Expressions," and I was hoping to clarify …
-
Given
```
schema doc {
document doc {
field chunks type array {}
}
field embedding type tensor(p{}, x[384]) {}
```
You can do things like
```
rank-profile example {
in…
-
## error log | 日志或报错信息 | ログ
Command:
```
pnnx best120.torchscript inputshape=[1,3,640,640]
```
Output:
```
pnnxparam = best120.torchscript.pnnx.param
pnnxbin = best120.torchscript.pnnx.bin
…
-
I am using einsum on four tensors, in a line like
`einsum(A, B, C, D)`
the shapes of A,B,C,D are, respectively: {3, 4, 4}, {3, 3, 2, 2}, {4, 2, 4}, {4, 2, 4}. This lines raises an error upon compila…