-
### Describe the bug
Not sure if its a bug. But the GPU usage blows up when using `kornia.morphology.erosion`.
For a tensor of shape `(1, 3, 958, 640)` & kernel of shape `(40, 40)`, the operatio…
-
I tried colorization from sketch using pix2pix model. However, sometimes there was a noise on the output picture. I slightly changed the code on data part: made a new file to read images from csv.
…
-
### 🐛 Describe the bug
I run the following code in a notebook:
```
import torch
import time
# Reserve 2G GPU RAM
mem_ptr = torch.cuda.caching_allocator_alloc(1024*1024*1024 * 2)
# Sleep 5…
-
To provide a specific PWM with only single signal input:
- [ ] add dead time with rising edge delay mode
- [ ] add dead time with balance mode rising edge delay and advance falling edge
- [ ] a…
-
[//]: # "SPDX-FileCopyrightText: Copyright (c) 2022-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved."
[//]: # "SPDX-License-Identifier: Apache-2.0"
[//]: # ""
[//]: # "Licensed under the …
-
### Describe the bug
While building an extension to (already fitted) GaussianMixture models (https://github.com/JohannesBuchner/askcarl/),
I was using https://hypothesis.readthedocs.io/ for testin…
-
I am opening this thread to coordinate contributions to new tutorials for use in the OpenMM 8 paper.
I'll keep this list updated with the discussion below:
- [ ] Extend `openmm-cookbook` to build …
-
### 🐛 Describe the bug
From a clean install using the current main branch, llama-3.2-11b-vision seems to need some love.
The download of the model files from HugginFace succeded using :
`python3 …
-
### 🐛 Describe the bug
Using ```torch.sigmoid``` on a tensor of negative complex numbers result in nan for cpu. For sigmoid, large negative numbers should result in 0, which it does in the cuda versi…
-
### Your current environment
```text
PyTorch version: 2.3.0+cu121
Is debug build: False
CUDA used to build PyTorch: 12.1
ROCM used to build PyTorch: N/A
OS: Ubuntu 22.04.3 LTS (x86_64)
GCC vers…