https://github.com/alan-turing-institute/ai-workflows
A collections of portable, real-world AI workflows for testing and benchmarking
https://github.com/alan-turing-institute/ai-workflows
ai apptainer hpc hut23 hut23-947 machine-learning singularity singularity-container
Last synced: 6 months ago
JSON representation
A collections of portable, real-world AI workflows for testing and benchmarking
- Host: GitHub
- URL: https://github.com/alan-turing-institute/ai-workflows
- Owner: alan-turing-institute
- License: mit
- Created: 2021-07-23T10:47:42.000Z (almost 4 years ago)
- Default Branch: main
- Last Pushed: 2022-07-21T09:49:37.000Z (almost 3 years ago)
- Last Synced: 2024-03-20T20:58:09.170Z (about 1 year ago)
- Topics: ai, apptainer, hpc, hut23, hut23-947, machine-learning, singularity, singularity-container
- Language: Shell
- Homepage:
- Size: 263 KB
- Stars: 8
- Watchers: 18
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# AI-workflows
A collections of portable, real-world AI workflows for testing and benchmarking
## Workflows
- [Pytorch GAN zoo](./workflows/pytorch_GAN_zoo/)
- [3D Very Deep VAE](./workflows/3d_very_deep_vae/)
- [SciML Bench](./workflows/sciml-bench/)
## Running Workflows on HPC
[hpc.md](./hpc.md) has advice on running the workflows on HPC systems. Each
workflow also contains template Slurm batch scripts.
## Using Vagrant
The included Vagrantfile can be used to build the examples if your system does
not support singularity or if you would prefer not to install singularity on
your host. The Vagrantfile has support for the VirtualBox and libvirt providers.
To create the virtual machine run
```bash
vagrant up --provider
```
for example
```bash
vagrant up --provider libvirt
```
When this is finished, you can connect to the virtualmachine in the normal way
```bash
vagrant ssh
```
and will find this project mounted at `/vagrant/`