Fabio Manganiello
cfedcd701e
Performance improvements when loading the Tensorflow plugin.
...
The Tensorflow module may take a few seconds to load the first time and
slow down the first scan of the plugins.
All the Tensorflow imports should therefore be placed close to where
they are used instead of being defined at the top of the module.
2023-05-11 19:48:22 +02:00
Fabio Manganiello
3bfc5b83ef
Moved to manifest files for describing plugins and backends and their dependencies
2021-09-16 17:53:40 +02:00
Fabio Manganiello
2a78f81a7b
Major LINT fixes
2021-04-05 00:58:44 +02:00
37e006d86e
Refined Tensorflow train methods
2020-10-12 01:06:32 +02:00
8d7e790eda
More descriptive attribute names for the Tensorflow predict response.
...
Using `outputs` instead of `values` and `predictions` instead of `labels`.
2020-10-02 01:17:34 +02:00
daaa0050d1
Squeeze the extra dimensions in a grayscale image and pass the model name to the response objects
2020-10-01 18:50:36 +02:00
287b6303ae
More flexible module loading and better lock management for models in Tensorflow plugin
2020-10-01 17:41:12 +02:00
3c3ee09d90
Fixed auto-generated docs indentation and code blocks
2020-08-31 17:39:42 +02:00
2e6388f6f4
ReadTheDocs fixes
2020-08-24 01:14:40 +02:00
7af8fb581b
A more robust logic to check whether the first dimension of the input tensor is None
2020-03-23 23:54:18 +01:00
375a8376dc
Don't use Config.get on __init__, or docs build will fail because of missing config.yaml
2020-03-23 01:14:44 +01:00
f4dcf688f0
Set default values for metrics for regression and networks
2020-03-23 01:10:59 +01:00
50e372be36
Implemented extensive support for neural networks, images and directories [ closes #121 ]
2020-03-23 01:00:25 +01:00
1f1fefca9d
Tensorflow plugin implementation WIP [ #121 ]
...
TODO: Extend neural network implementation to work
also with e.g. input from images, sounds or binary
2020-03-21 19:50:59 +01:00
Fabio Manganiello
Fabio Manganiello