-
## 🐛 Bug
With Visual Studio 2019, build from source (directly invoke CMakeLists) with `USE_VULKAN ON` (and `USE_VULKAN_SHADERC_RUNTIME ON`) fails due to usage of C++20 features in libtorch sour…
-
**Describe the bug**
Using `torch.compile` to optimize MONAI transforms generally works (apart from graph breaks), but `CenterSpatialCrop` (and its dictionary wrapper) does not.
**To Reproduce**
…
-
As i am trying to load custom pytorch lstm model in android studio using torch.load() .. it is throwing error which is mentioned below.
Would anyone mind helping me out in this. What is the right way…
-
Got the error during running the train.py file:
Transferred 308/361 items from yolov5s.pt
Scaled weight_decay = 0.0005
optimizer: Adam with parameter groups 59 weight (no decay), 62 weight, 62…
-
Thanks for propose a caffe2 android demo, But it can't pass the build and compile step when this project passed to new verison. It always has a link problem. So I ask for the previous version which in…
-
I'm building caffe2 on ubuntu 18.04 with CUDA 9.0? But when I run "python setup.py install" command, I have met issue about version of CUDA. It needs to CUDA 9.2 instead of 9.0 but i only want to buil…
-
## 🐛 Bug
I am trying to create an FP16 tensor in caffe2, and I am following the method described in https://github.com/pytorch/pytorch/blob/master/caffe2/proto/caffe2.proto#L86 (set type to FLOAT1…
-
Rencently, I need depthwise separable convolution for my project in order to reduce the parameters.
Does the caffe2 have the depthwise separable convolution oprater ?
-
GPU test passed. However, when I run code no PID is shown on nvidia-smi. Is there other way to check whether the code is running on GPU? Or is there some more lines of code I need to add so that caffe…
-
Here is the full set of warnings from a recent run:
```
Max No attributes
Shape No attributes=============================== warnings summary ===============…