Notes on Tensorflow install

master
ml server 2020-01-15 17:06:04 -07:00
parent 14fc5083c6
commit 787ca1e226
1 changed files with 30 additions and 16 deletions

View File

@ -79,33 +79,47 @@ apt update
apt install curl jq
```
## Machine Learning
For the machine learning scripts, like `wut-ml`, both Tensorflow
and Keras need to be installed. The versions of those in Debian
didn't work for me. IIRC, for Tensorflow I built a `pip` of
version 2.0.0 from git and installed that. I installed Keras
with `pip`. Something like:
## Install Tensorflow
For the machine learning scripts, like `wut-ml`, Tensorflow
needs to be installed.
As of version 2 of Tensorflow, Keras no longer needs to be
installed separately.
The verions of Tensorflow installed with `pip3` on Debian
Buster crashes. It is perhaps best to do a custom install,
based preferred build options, of the most preferred version.
At this point, the `remotes/origin/r2.1` branch is preferred.
To install Tensorflow:
1. Install dependencies in Debian.
1. Install Bazel to build Tensorflow.
1. Build Tensorflow pip package.
1. Install Tensorflow from custom pip package.
```
# XXX These aren't the exact commands, need to check...
# Install deps
apt update
# deps...
apt install python3-pip ...
# Install bazel or whatever their build system is
apt install python3-pip
# Install bazel .deb from releases here:
firefox https://github.com/bazelbuild/bazel/releases
# Install Tensorflow
git clone tensorflow...
cd tensorflow
git checkout remotes/origin/r2.1
./configure
# run some bazel command
dpkg -i /tmp/pkg_foo/*.deb
apt update
apt -f install
# Install Keras
pip3 install --user keras
pip3 install --user /tmp/pkg_foo/*.whl
# A million other commands....
```
# Usage
The main purpose of the script is to evaluate an observation,
but to do that, it needs to build a corpus of observations to