Welcome to Salamander!
This guide takes about 1 minute to complete. Once complete, you will have access to a GPU-enabled Jupyter Notebook and the course-v3 materials
If you are returning to work and have previously completed the steps below, please go to the returning to work section.
We recommend the “Accelerated Computing” configuration which costs $0.45 per hour.
You should use the suggested 75GB disk size, which is an additional $7.50 per month. You can increase the disk size later if you need more space.
Free credits for students
You can redeem any AWS coupon on Salamander. Students can claim between $75 to $110 free credits via the GitHub education pack.
How much will you use this course
Considering that the course requires, over 2 months, 80 hours of homework plus the 2 hours of working through each lesson, we calculated roughly how much you would spend in the course.
- Accelerated Computing + Storage: (80+2*7)*$0.38 + $7.50*2 = $50.72
Step 1: Create an account
Visit https://salamander.ai, click “Get Started”, complete the signup form, and add your card details.
Students can claim their free credits at https://salamander.ai/redeem-aws-coupon.
Step 2: Create your server
Make sure the “PyTorch 1.0, fastai, and the v3 MOOC course” software is selected, leave the remaining defaults as they are, and click “Launch server”.
Your server will appear straight away, and after a couple minutes the orange status indicator will turn blue.
Step 3: Open Jupyter Notebook
Click ‘Jupyter Notebook’ to access the course materials. After Jupyter Notebook loads, click on ‘course-v3’.
See here for instructions on running the Jupyter Notebook tutorial. Return to this guide once you’re finished and don’t forget to stop your server.
Step 4: Update the fastai library
Before you start working it’s a good idea ot update the pytorch and fastai libraries, and the course notebooks. To do this you will have to access the terminal. You can do this by clicking in ‘New’, ‘Terminal’.
Once you click on ‘Terminal’ a new window should open with a terminal. Type (or copy and paste):
conda install -c fastai -c pytorch fastai pytorch cd course-v3/ git pull
These steps may take a couple of minutes to complete. When done, you should close the browser tab containing the terminal. You can now click on “course-v3”, and then the first course notebook, and get started!
Step 5: Stop your server
When you’re all done, don’t forget to shut down your server, so you don’t get charged for the time it’s running in the background. It’s not enough to just close your browser or turn off your own computer. Go back to salamander and click the ‘Stop Server’ button next to your server.
For more details, & updating the course / fastai library see “Returning to work”.
Advanced users: Full server access
Salamander isn’t just for Jupyter Notebook - each server is a full Ubuntu instance. To connect, click “Setup Access Key” and generate or upload an ssh key - it’ll get added to all of your servers automatically. You can then copy & paste
ssh ubuntu@[xxx.xxx.xxx.xxx] from the webpage to your terminal. Press enter and you’re in!
Note: you should always generate keys yourself if you’d like to use them for several different platforms (in-depth guide)
Many thanks to Ashton Six for writing this guide.