You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+12-12Lines changed: 12 additions & 12 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -2,7 +2,7 @@
2
2
3
3
This is my attempt to build a recent Tensorflow wheel and Docker image compatible with multiple architectures and processors.
4
4
5
-
It's currently using Tensorflow v2.14.0
5
+
It's currently using Tensorflow v2.16.1
6
6
7
7
It supports the following architectures:
8
8
- ARM32
@@ -16,7 +16,7 @@ The docker images are `docker.io/snowzach/tensorflow-multiarch` [here](https://h
16
16
The default/latest docker images includes arm32, arm64 and amd64-noavx
17
17
18
18
## Compiling Tensorflow Wheel
19
-
I checked out the Tensorflow source at version 2.14.0 and I had to edit the file `tensorflow/tools/ci_build/pi/build_raspberry_pi.sh` and insert on line 99 to add an ARMHF target.
19
+
I checked out the Tensorflow source at version 2.16.1 and I had to edit the file `tensorflow/tools/ci_build/pi/build_raspberry_pi.sh` and insert on line 99 to add an ARMHF target.
After running the command there was working python whl and library in the `output-artifacts` directory.
37
37
38
-
I renamed the armhf output wheel from `tensorflow-2.14.0-cp38-none-linux_armhf.whl` to `tensorflow-2.14.0-cp38-none-linux_armv7l.whl` to ensure it matched the hardware architecture and would be installed by python.
38
+
I renamed the armhf output wheel from `tensorflow-2.16.1-cp38-none-linux_armhf.whl` to `tensorflow-2.16.1-cp38-none-linux_armv7l.whl` to ensure it matched the hardware architecture and would be installed by python.
39
39
40
40
# Tensorflow IO
41
41
Tensorflow IO is required in order to install the Tensorflow wheel. There seems to be a chicken-egg problem
0 commit comments