-
## Issue description
I found a few bugs in NNApi integration while running benchmarks.
## Code example
Bug 1: Invalid size is passed for tensors.
Caffe2 seems to be passing invalid size for tenso…
-
Hello,
While compiling generated sources(caffe2.pb.cc) from caffe2.proto we are getting the implicit conversion error
protobuf version=v3.5.2
compiler:
CC=aarch64-linux-android-clang \
CXX=a…
-
If you have a question or would like help and support, please ask at our
[forums](https://discuss.pytorch.org/).
If you are submitting a feature request, please preface the title with [feature req…
-
Hi
I install Caffe2 with Python3.6.8, cuda 9.0.176, cudnn 7 ,Driver Version: 384.130 OS ubuntu 17.10
Caffe2 install use: conda install pytorch-nightly cuda90 -c pytorch
when I try to run mask…
-
## 🐛 Bug
I want to use the `caffe2::Predictor` API, but when I download `libtorch` from the website, even on the current stable (1.0) zip file, `libtorch/include/caffe2/predictor` folder is entirel…
-
## 📚 Documentation
First I tried installing using pip into a Python 3.7 environment. That mostly worked, except the Loading_Pretrained_Models tutorials failed. Apparently it uses a python2 module. Ma…
-
I compile caffe2 from source(pytorch) by anaconda3.
operations:
In the pytorch directory,I
mkdir build && cd build
cmake -DCMAKE_PREFIX_PATH=/home/zhengr/.conda/envs/caffe2-source -DCMAKE_INS…
-
I noticed that when I run my caffe2 model in c++, it only utilizes a single core.
In tensorflow C++ API, I had a choice on the specific number of threads/cores I want to use to run the network. Is …
-
## 🚀 Feature
The conv op has the group attribute, but it's not in convTranspose.
## Motivation
The convTranspose should have the same functionality as conv. And pytorch convT op has group a…
-
Hi,
the conversion of the model to caffe2 via ONNX does not work with Pytorch 1.0.1. The initial script converts the model to ONNX, but the conversion to caffe2 fails, as
`init_net, predict_net =…