Robin

Hi, It appears I have the same issue here, but my bazel version is the 0.4.3 I'm on 16.04, TF without the GPU. I'm on deepdetect master with commit :...

Hi @beniz, I run the following cmake commande : `cmake .. -DUSE_XGBOOST=0N -DUSE_TF=ON -DUSE_TF_CPU_ONLY=ON` The first issue I encountered was that some TF files (such as image_ops.h) were left in...

@beniz I still encounter this issue. the result of the "bazel version" command gives me this : ``` Build label: 0.4.3 Build target: bazel-out/local-fastbuild/bin/src/main/java/com/google/devtools/build/lib/bazel/BazelServer_deploy.jar Build time: Thu Dec 22 12:31:25...

Huh, okay that last mistake is on me, my LD_LIBRARY_PATH was defined as /usr/local/cuda-8.0/ and not cuda, as my python tensorflow is using this one. Ok so, It build now....

Okay. So to solve this, and as stated in my previous comment : I start like in the readme : ``` git clone https://github.com/beniz/deepdetect cd deepdetect mkdir build cd build...

Got the same issue, solved it by rolling back to kedro 16.2.

Anyone has a solution for the "vectors$vectorid is not an integer vector" issue?