-
### 🐛 Describe the bug
I use a variety of transforms for my training, however the `grid_sampler` seems to be not implemented for cpu when using mixed training on fp16? Is there any solution to use al…
-
Can LegacyUpdate ports latest security and most important features that most softwares depends on it (how the future goes...) from latest Windows to older versions of Windows and push it as an update …
-
@soulitzer, @bdhirsh, @ezyang if I missed any:
Tracking here so I don't need to keep all of them in my head
- [ ] way to mark operator as non-differentiable
- [x] ctx.needs_input_grad is wrong fo…
-
Setup :
- `This is GNU Emacs 28.2 (build 1, x86_64-pc-linux-gnu, GTK+ Version 3.24.37, cairo version 1.16.0) of 2023-03-14, modified by Debian`
- from `packages-list-package` : `jupyter 20230214.…
-
I recently updated from 36010 to 36600 and facing boot issues after that on my laptop. Luckily I have multiple kernels installed. I can boot from **org.clearlinux.lts2019.5.4.196-188** and **org.clear…
-
### Describe the issue
When using the WebGL backend with [this model](https://huggingface.co/rocca/simvp-web/resolve/main/onnx/moving_mnist/epoch-137-quantized.onnx), I get the following error:
```
…
-
### What happened?
I've compiled the latest build of `llama-server` (version b3205) using the method recommended in the docs.
The recent renaming of `server` to `llama-server` has tempted me to us…
-
**Bug summary**
My understanding is that `sycl::native::` functions are supposed to be faster, less precise versions of the "default" `sycl::` math functions. However, it seems that (at least with cl…
-
https://jax.readthedocs.io/en/latest/pallas/design.html
Create pallas kernel for LIF neurons and investigate to see if it's faster than just static unrolling the LIF as defined in Spyx.
-
I'm trying to render glass with caustics. When I select CPU, it processes apparently normally, i.e. the processor gets at full power and the samples counter adds up at some samples/second. But I get o…