[WIP] Machine-Learning-Training

Work in progress

Welcome to this repository

You might ask yourself what can I find here. This repository mainly contains different small and large projects related to machine learning.

Contains jupyter notebooks and examples for Keras, Google AI Platform and Kubeflow.

I update this repository while learning new technologies and topics. It is used as a self reminder for my future self.

Repository structure

.
├── ai-platform                     # Implementations related to Google AI Platform.
├── kubeflow                        # Implementations related to kubeflow.
    └── leightweight-component-training # Explains the basics of components and pipelines in Kubeflow
├── notebooks                       # Several notebooks for different kind of technologies.
├── notes                           # Personal machine learning notes, as self reminder =).
└── presentations                   # Personal presentations related to ml.

Changelog

v0.5

2019-10-24

• added TFRecords with tf.hub feature extraction sample

v0.4

2019-06-26

• added training component for concatenating strings
• added training notebook for reusable components and pipelines

v0.3

2019-06-18

• NER added deployment for AI Platform using beta feature for custom processing routine

v0.2

2019-06-17

• NER added processing state export for custom prediction routine usage
• NER added logic to export keras model

v0.1

2019-06-14

• Initial repository and structure.
• Added several basic kubeflow notebooks.