Generic Framework for ML projects
Project description
CoreML
Generic framework for Machine Learning projects.
Setup
Clone the project:
$ git clone https://github.com/dalmia/coreml.git
Weights & Biases
We use wandb
for experiment tracking. You'll need to have that set up:
- Install wandb
$ pip install wandb
- Login to wandb:
$ wandb login
You will be redirected to a link that will show you your WANDB_API_KEY .
- Set the WANDB_API_KEY by adding this to your ~/.bashrc file:
export WANDB_API_KEY=YOUR_API_KEY
- Run
source ~/.bashrc
.
Docker
We use Docker containers to ensure replicability of experiments. You can either fetch the Docker image from DockerHub using the following line:
$ docker pull adalmia/coreml:v1.0
OR
You can build the image using the DockerFile
:
$ docker build -t adalmia/coreml:v1.0 .
The repository runs inside a Docker container. When creating the container, you need to mount the directory containing data to /data
and directory where you want to store the ouptuts to /output
on the container. Make the corresponding changes to create_container.sh
to mount the respective directories by changing /path/to/coreml
, /path/to/data
and /path/to/outputs
to the appropriate values.
Use the following command to launch a container:
$ bash create_container.sh
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.