Installation
Ersilia Model Hub installation instructions for local inference
1. Install on Windows
We are not testing Windows installation consistently. If you encounter problems, please reach out to us at hello@ersilia.io.
We recommend that you install Ubuntu on your Windows machine. This can be now done very easily with WSL. You will need at least Windows 10.
Open a Power Shell with Admin permissions and type:
Then simply install the Ubuntu terminal on Windows. Inside the Ubuntu terminal, you can now follow the installation instructions for Linux below.
2. Install on Linux and MacOSX
2.1. Pre-requisites
There are a few third-party pre-requisites and you need to have them installed on your computer before proceeding with the installation of the Ersilia tool. Don't worry, these dependencies are very popular, well documented and continuously maintained. Most probably you know them already.
If you already meet some of the pre-requisites you may skip the corresponding steps.
Pre-requisite 1: The gcc compiler
Probably you have the gcc compiler installed already. This is the command to install it in Ubuntu (the command may be different if you do not use Ubuntu):
Pre-requisite 2: Python and Conda
We have tested our tools on Python 3.8 and above. Please make sure you have the right Python installation on your computer. Visit the official Python site to learn more.
Although not strictly necessary, we encourage the use of Conda environments. If Conda is available in the local device, the Ersilia tool will try to use it in order to handle dependencies safely and efficiently. Conda can be installed following these instructions. Many people choose to use Anaconda or Miniconda installers.
If you don't have the Conda package manager yet, we suggest you install Miniconda. This is the command to install it in Ubuntu (the command may be different if you do not use Ubuntu):
Pre-requisite 3: Git and GitHub CLI
All of our code is hosted in the Ersilia GitHub Organization profile. The Ersilia tool needs to fetch code from GitHub, so make sure Git is installed on your device. You can follow these instructions.
In addition, we highly recommend the fantastic GitHub CLI, which is used by Ersilia if available.
Once Conda is installed (see above), you can use it to install GitHub CLI
Use the GitHub CLI to login. This may be helpful if you have contributor permissions at Ersilia. Type:
And then follow the instructions.
Pre-requisite 4: Git LFS
Some pre-trained models, especially the ones that contain many parameters, require Git Large File Storage (LFS).
Git LFS can be installed from Conda:
Activate Git LFS:
Pre-requisite 6: Docker
Docker containers are an excellent way to share applications and simplify the management of system requirements and configurations. Please install Docker to ensure that all our AI/ML assets will work smoothly in your local device.
2.2. Install Ersilia
After all the pre-requisites are met, we are ready to install the Ersilia tool.
Open a terminal and set up a Conda environment.
Then, simply install the Ersilia Python package.
You should be done! Check that the CLI works on your terminal, and explore the available commands
The Ersilia Model Hub is growing continuously to fulfill the needs of the community. Please do not hesitate to request new models!
[DEPRECATED] Extra-requisite: The Isaura data lake
DEPRECATION NOTICE: Please note that Isaura is no longer an actively maintained project. While we are working on fixing that, current versions of Ersilia break if Isaura is installed in the same environment. We recommend not installing Isaura at this point of time.
We highly recommend installation of the Isaura data lake. With Isaura, you will be able to cache your model predictions (i.e. store them in your local computer). Isaura is a relatively light Python package:
Last updated