Tensorflow 2.1.0

# md5sum
10a46e8ee9c0b27437a26d7a4252d452  tensorflow-2.1.0-cp37-cp37m-linux_x86_64_with_SSE3_SSE4.1.whl

Tensorflow has been built following the documentation on Debian with the following CPU :

Architecture:        x86_64
CPU op-mode(s):      32-bit, 64-bit
Byte Order:          Little Endian
Address sizes:       40 bits physical, 48 bits virtual
CPU(s):              18
On-line CPU(s) list: 0-17
Thread(s) per core:  1
Core(s) per socket:  6
Socket(s):           3
NUMA node(s):        1
Vendor ID:           GenuineIntel
CPU family:          6
Model:               29
Model name:          Intel(R) Xeon(R) CPU           X7460  @ 2.66GHz
Stepping:            1
CPU MHz:             2659.778
BogoMIPS:            5319.55
Hypervisor vendor:   KVM
Virtualization type: full
L1d cache:           32K
L1i cache:           32K
L2 cache:            4096K
L3 cache:            16384K
NUMA node0 CPU(s):   0-17
Flags:               fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx lm constant_tsc arch_perfmon rep_good nopl xtopology cpuid tsc_known_freq pni ssse3 cx16 sse4_1 x2apic tsc_deadline_timer hypervisor lahf_lm cpuid_fault pti tsc_adjust arat arch_capabilities

How to build

# On Debian
apt install python3-dev python3-pip git screen
pip3 install --upgrade pip
pip3 install -U --user pip six numpy wheel setuptools mock 'future>=0.17.1'
pip3 install -U --user keras_applications --no-deps
pip3 install -U --user keras_preprocessing --no-deps

You need to download Bazelisk. You can add the directory where you downloaded Bazelisk in your PATH temporarily.

export PATH=/path/to/folder:$PATH

Then you need to git clone Tensorflow :

git clone https://github.com/tensorflow/tensorflow.git
cd tensorflow
# If you want to build a specific version on your computer
# git checkout branchname 

Then you need to configure the system build. You can let everything by default EXCEPT the location of Python (I needed to set /usr/bin/python3 on my Debian) :

./configure

Finally you are ready to build.

Tips : As I built Tensorflow on a server, I run the script thanks to screen :

screen
bazel build --config=opt //tensorflow/tools/pip_package:build_pip_package

You can now grab a coffee (or several, because it's very long... More than 10hours on my (old) computer).

Finally, you can build the package and install it :

./bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg
pip install /tmp/tensorflow_pkg/tensorflow-*.whl