how to make tensorflow-gpu for px2

Autonomous Vehicles DRIVE PX2 General
#1

hi

i really need a tensorflow-gpu 1.9 for px2 run with python 3.5 for a unet, i really dont see any tensorflow-gpu version for aarch64, and I cant find the tensorflow-gpu source code(only found tensorflow source code). and i really need cmake config,

i have a tensorflow that run with cpu which’s very slow. i really need it run with dGPU.

thanks all
hanbinli

#2

Hi hanbinl,

We don’t have a prebuilt for python3.5 on PX2.
But you can follow the steps to build a python3 version on your own.
https://devtalk.nvidia.com/default/topic/1049100/general/tensorflow-installation-on-drive-px2-/post/5324624/#5324624

See also: https://devtalk.nvidia.com/default/topic/1051675/general/installation-help-needed-for-tensorflow-in-python-3-5-drive-px2/

Thanks.

#3

hi kayccc,

i have a problem with java8 install by using sudo apt-get install -y oracle-java8-installer

which shows:

sudo apt-get install -y oracle-java8-installer 
[sudo] password for nvidia: 
Reading package lists... Done
Building dependency tree       
Reading state information... Done
Package oracle-java8-installer is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source

E: Package 'oracle-java8-installer' has no installation candidate

is any way to deal with?

#4

Hi

I can’t answer that due to no experience on oracle-java8-installer, but please refer to below topic if can help:
https://devtalk.nvidia.com/default/topic/1052181/jetson-nano/java-on-jetson-nano/post/5342530/#5342530

Thanks

#5

hi kayccc,

i have followed the steps, and i got a question that how to change the following in step 4

diff --git a/tensorflow/lite/kernels/internal/BUILD b/tensorflow/lite/kernels/internal/BUILD
index 4be3226..960754a 100644
--- a/tensorflow/lite/kernels/internal/BUILD
+++ b/tensorflow/lite/kernels/internal/BUILD
@@ -22,7 +22,6 @@ HARD_FP_FLAGS_IF_APPLICABLE = select({
 NEON_FLAGS_IF_APPLICABLE = select({
     ":arm": [
         "-O3",
-        "-mfpu=neon",
     ],
     ":armeabi-v7a": [
         "-O3",
diff --git a/third_party/aws/BUILD.bazel b/third_party/aws/BUILD.bazel
index 5426f79..e08f8fc 100644
--- a/third_party/aws/BUILD.bazel
+++ b/third_party/aws/BUILD.bazel
@@ -24,7 +24,7 @@ cc_library(
         "@org_tensorflow//tensorflow:raspberry_pi_armeabi": glob([
             "aws-cpp-sdk-core/source/platform/linux-shared/*.cpp",
         ]),
-        "//conditions:default": [],
+        "//conditions:default": glob(["aws-cpp-sdk-core/source/platform/linux-shared/*.cpp",]),
     }) + glob([
         "aws-cpp-sdk-core/include/**/*.h",
         "aws-cpp-sdk-core/source/*.cpp",
diff --git a/third_party/gpus/crosstool/BUILD.tpl b/third_party/gpus/crosstool/BUILD.tpl
index db76306..184cd35 100644
--- a/third_party/gpus/crosstool/BUILD.tpl
+++ b/third_party/gpus/crosstool/BUILD.tpl
@@ -24,6 +24,7 @@ cc_toolchain_suite(
         "x64_windows|msvc-cl": ":cc-compiler-windows",
         "x64_windows": ":cc-compiler-windows",
         "arm": ":cc-compiler-local",
+        "aarch64": ":cc-compiler-local",
         "k8": ":cc-compiler-local",
         "piii": ":cc-compiler-local",
         "ppc": ":cc-compiler-local",

thanks