search

EU-TEACHING / teaching-app

Repository for TEACHING application startup and deployment. This can be taken as the main access point for the TEACHING plaform developed within the EU TEACHING 2020 project.
MIT License
17 stars 4 forks source link

Integration of the AI-Toolkit with the TEACHING_Platform #3

Closed antodima closed 2 years ago

antodima commented 2 years ago

Hi all! :wave::wave: I have created another scenario in order to integrate the AI-Toolkit with the TEACHING_Platform. I added the following files:

  1. Dockerfile.ai_toolkit: the Docker image that clones the AI-Toolkit repository and copies all the necessary files to instantiate a specific Learning Module (at the moment it is capable of instantiate only an Echo State Network (ESN) for stress prediction)
  2. scenario_5.yaml: the Docker-compose file with rabbitmq, file_multisensor and ai_toolkit images
  3. eda.csv into "data_storage/raw_data/wesad/" folder: is the dataset generated from wesad_s2_empatica.pkl file. It is passed to the file_multisensor image by the FILE_SOURCE environment variable into scenario_5.yaml in order to stream the EDA values used by the ESN Learning Module

:warning: In order to have a correct run of the scenario_5 it is important to set the GITHUB_ACCESS_TOKEN variable in the scenario_5.yaml under ai_toolkit > build > args. I generate the token from Settings > Developer settings > Personal access tokens of the Github profile.