Vessl Docs
Search…
Quickstart
Run your first experiment on Vessl
Follow along with the examples from Vessl's GitHub repository: https://github.com/vessl-ai/examples

1. Sign up and create an Organization

To begin your machine learning project with Vessl, sign-up at https://vessl.ai/ and create an Organization. Organization is where your teammates can find shared assets such as datasets, models, and experiments.

2. Create a Project

Fork the examples from Vessl's GitHub repository to your GitHub account. Then, create a Version control project. Import the GitHub repository titled examples on the dropdown menu and and fill in the name and description of the project.

3. Install Vessl Client

Install Vessl Client on your local machine using pip install. Vessl Client is a unified tool to manage machine learning workflow and assets on Vessl. It includes:
    Command Line Interface (CLI) used to handle datasets and experiments.
    Python SDK used for logging experiment metrics.
1
pip install savvihub
Copied!

4. Git clone a repository and login to Vessl Client

Git clone Vessl's GitHub repository and login with Vessl Client.
1
git clone https://github.com/{your-github-account}/examples.git
2
cd examples
3
savvihub login
Copied!

5. Run an experiment

Run experiment command on your local machine either with command options or inquirers.
Run with Options
Run with Inquiries
1
sv experiment run \
2
-c aws-apne2-prod1 \
3
-r c2.r6.spot \
4
-i savvihub/kernels:py36.full-cpu \
5
--start-command 'pip install -r mnist/requirements.txt && python mnist/keras/main.py --save-model --save-image'
Copied!
1
sv experiment run
2
[?] Start command: pip install savvihub && python mnist/keras/main.py --save-model
3
[?] Please choose a cluster: [1] aws-apne2-prod1 (SavviHub)
4
> [1] aws-apne2-prod1 (SavviHub)
5
6
[?] Please choose a resource: [1] c2.r6.spot (CPU 2 Cores / Memory 6GB)
7
> [1] c2.r6.spot (CPU 2 Cores / Memory 6GB)
8
[2] c2.r6 (CPU 2 Cores / Memory 6GB)
9
[3] c4.r13 (CPU 4 Cores / Memory 13GB)
10
[4] g1a.r6 (GPU(K80) x 1 / CPU 2 Cores / Memory 6GB)
11
[5] g2a.r13 (GPU(K80) x 2 / CPU 4 Cores / Memory 13GB)
12
[6] g1b.r13 (GPU(V100) x 1 / CPU 4 Cores / Memory 13GB)
13
[7] g2b.r13 (GPU(V100) x 2 / CPU 4 Cores / Memory 13GB)
14
[8] c0.r1 (CPU shared / Memory 1GB)
15
16
[?] Please choose a kernel image: [1] savvihub/kernels:py36.full-cpu (Python 3.6 (All Packages))
17
> [1] savvihub/kernels:py36.full-cpu (Python 3.6 (All Packages))
18
[2] savvihub/kernels:py37.full-cpu (Python 3.7 (All Packages))
19
[3] savvihub/kernels:py36.full-cpu.jupyter (Python 3.6 (JupyterLab))
20
[4] savvihub/kernels:py37.full-cpu.jupyter (Python 3.7 (JupyterLab))
21
[5] tensorflow/tensorflow:1.14.0-py3 (Tensorflow 1.14.0)
22
[6] tensorflow/tensorflow:1.15.5-py3 (Tensorflow 1.15.5)
23
[7] tensorflow/tensorflow:2.0.4-py3 (Tensorflow 2.0.4)
24
[8] tensorflow/tensorflow:2.2.1-py3 (Tensorflow 2.2.1)
25
[9] tensorflow/tensorflow:2.3.2 (Tensorflow 2.3.2)
26
[10] tensorflow/tensorflow:2.4.1 (Tensorflow 2.4.1)
27
[11] tensorflow/tensorflow:2.3.0 (TensorFlow 2.3.0 (Tensorboard))
28
29
Run experiment with revision 8a6f65c (HEAD)
30
Commit: 8a6f65c Update keras mnist example (2 minutes ago) <ryooit>
31
32
Experiment 1 is running. Check the experiment status at below link
33
https://savvihub.com/floyd/examples/experiments/1
Copied!

6. Monitor your experiment with Vessl Web Console

Congratulations! You ran your first experiment on Vessl! Visit Vessl Web Console to see your experiment results.
Last modified 15d ago