Anton Lokhmotov

This is a bit tricky for a couple of reasons: - TensorRT 3.0 needs CUDA 9.0; presumably, TensorFlow should use the same CUDA version. - TensorRT 3.0 needs cuDNN 7.0;...

To configure TensorRT 3 to support Python 2, I followed the official installation guide. ## Install and detect TensorRT 3 ``` $ cd /tmp $ tar xvzf ~/Downloads/TensorRT-3.0.4.Ubuntu-16.04.3.x86_64.cuda-9.0.cudnn7.0.tar.gz $ sudo...

From the user perspective, it seems that the easiest way of installing TensorRT with Python support would be to download an archive from NVIDIA's website (registration is required) and then...

Compiling with Clang 6 from Android NDK r17b results in even more errors and warnings: ``` clang++ -c -fPIE -pie -target aarch64-linux-android -gcc-toolchain /home/anton/data/android-ndk-r13b/toolchains/aarch64-linux-android-4.9/prebuilt/linux-x86_64 --sysroot=/home/anton/data/android-ndk-r13$ /platforms/android-24/arch-arm64 -I/home/anton/data/android-ndk-r17b/sysroot/usr/include -I/home/anton/data/android-ndk-r17b/sysroot/usr/include/aarch64-linux-android -I../ -DCK_HOST_OS_NAME2_LINUX=1...

Similar for NDK 13b: ``` $ l /home/anton/data/android-ndk-r13b/toolchains/llvm/prebuilt/linux-x86_64/include ls: cannot access '/home/anton/data/android-ndk-r13b/toolchains/llvm/prebuilt/linux-x86_64/include': No such file or directory $ find /home/anton/data/android-ndk-r13b/toolchains/llvm/prebuilt/linux-x86_64/ -name include /home/anton/data/android-ndk-r13b/toolchains/llvm/prebuilt/linux-x86_64/lib64/clang/3.8/include /home/anton/data/android-ndk-r13b/toolchains/llvm/prebuilt/linux-x86_64/lib64/clang/3.8.256229/include ```

For Clang 7 in NDK 18b: ``` $ ck show env --tags=compiler,lang-cpp,llvm,target-os-android28-arm64,v7 Env UID: Target OS: Bits: Name: Version: Tags: 1c70f43a4910bf0c android28-arm64 64 Android NDK LLVM compiler 7.0.2 64bits,android,compiler,host-os-linux-64,lang-c,lang-cpp,llvm,ndk,target-os-android28-arm64,v7,v7.0,v7.0.2 ```...

We now have quite a few Docker images for TFLite, ArmNN-TFLite and TF-C++: ``` anton@velociti$ docker image ls REPOSITORY TAG IMAGE ID CREATED SIZE ctuning/object-detection-armnn-tflite.debian-9 latest 8d8e41f08acc 6 minutes ago...

One of the issues with Docker images (however incredibly helpful they are in expanding our testing capability) is the sheer size of them, especially when there contents is not reused....

Still, it would nice to consider creating "ensembles" of CK Docker images. Perhaps we could place CK components (libraries, datasets) on virtual drives, and seamlessly map them from inside Docker...

For example, for `ctuning/object-detection-armnn-tflite.debian-9` and `ctuning/object-detection-tflite.debian-9` this "Venn Diagram intersection" would contain the TFLite library and the non-quantized MobileNet model. On the other hand, `armnn-tflite` uses only the first 50...