-
I try to run resnet50_trainer.py to build my model by using 100 images, but I get the following error messages. My train method is drawn on the following web, https://github.com/royhuang9/Detectron. H…
-
I modified CMakelist.txt to turn off USE_OPENMP. I change option(USE_OPENMP "Use OpenMP for parallel code" ON) into (USE_OPENMP "Use OpenMP for parallel code" OFF). Then I got compilation error as be…
-
### 🐛 Describe the bug
When using LibTorch in CMake via `find_package(Torch REQUIRED)` it tries to find CUDA via `nvcc`. When only the runtime libraries are installed, e.g. `sudo apt install cuda-lib…
-
Hello, thank you for sharing. I always get an error at caffe2 when compiling libtorch. I've been looking for a long time and still haven't solved the problem. It's probably CUDN and CUDNN. So, does yo…
-
- [x] Add
- [x] Div
- [x] Mul
- [ ] Pow
- [x] Sub
- [ ] And
- [ ] Or
- [ ] Xor
- [x] Equal
- [x] Greater
- [x] Less
- [ ] Gemm
- [ ] PRelu
-
## 🐛 Bug
First, this is not actually a bug, but when one uses pytorch to train a model and then convert the model to onnx which later will be used to do inference by Caffe2, then different result…
-
Hey Caffe2/PyTorch team,
I had a question about going from a trained model (.mdl format) to a standalone predictor model. I've been able to load the .mdl file as follows:
```
meta_net_def = pred_…
-
## 🚀 Feature
Similar to the Pytorch implementation of [crossentropyloss](https://pytorch.org/docs/stable/nn.html#crossentropyloss), it'd be nice to have an `ignore_index` or `ignore_label` argument…
-
ubuntu@ubuntu-Z270N-WIFI:~/caffe2$ make && cd build && sudo make install
-- Setting CMAKE_FIND_NO_INSTALL_PREFIX
-- Using protobuf compiler /usr/bin/protoc
-- The BLAS backend of choice:Eigen
-- C…
-
**Describe the bug**
When I want to convert the onnx model to the tf model, a unique operator does not support it. Is there anything I can replace this operator?
**To Reproduce**
Please give …