https://github.com/dominodatalab/reference-project-mnist-digits

Build a CNN. Deploy a model API. Build and deploy a shiny app to access the model API.
https://github.com/dominodatalab/reference-project-mnist-digits

Last synced: 4 months ago
JSON representation

Build a CNN. Deploy a model API. Build and deploy a shiny app to access the model API.

• Host: GitHub
• URL: https://github.com/dominodatalab/reference-project-mnist-digits
• Owner: dominodatalab
• Created: 2021-10-07T22:01:51.000Z (almost 4 years ago)
• Default Branch: master
• Last Pushed: 2021-10-12T20:10:47.000Z (almost 4 years ago)
• Last Synced: 2025-01-24T18:27:45.338Z (6 months ago)
• Language: Jupyter Notebook
• Size: 12.9 MB
• Stars: 0
• Watchers: 6
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README

Awesome Lists containing this project

README

        
## Building and Deploying a CNN

![image](raw/latest/scratch/domino-nvidia-aws.png)

1. Train a CNN model using a Tesla V100 GPU and a Domino environment preconfigured with an NVIDIA driver, CUDA 9, Tensorflow, and Keras. 

    * Use this notebook to train and save the model: [MNIST_keras.ipynb](view/MNIST_keras.ipynb)

    

2. Use Domino to deploy and manage a model API Endpoint on AWS. 

    * Use this file as when building the API: [MNIST_predict_list.py](view/MNIST_predict_list.py)

    

3. Use Domino to deploy and manage a Web App on AWS allowing others to interact with your model.

    * Use these files to define the web app: [ui.R](view/ui.R) and [server.R](view/server.R)

    

![image](raw/latest/mnist.png)

app inspiration taken from https://github.com/agenis/handwritten-digits