PyTorch Lab for the BMVA Summer School
Please read the PDF for the introduction to PyTorch.
Use Google colaboratory to run the notebooks!
http://colab.research.google.com/
-
Open a browser and go to http://colab.research.google.com/
-
Select "Open Notebook" from the menu
- Select "GitHub" and type "
ndfcampbell/bmva_summer_school" into the address bar
- Load the PyTorch_Example file first and then take a look at the linear regression lab!
- If you are curious as to how PyTorch (and similar libraries) work behind the scenes then please take a look at the Scalar Auto Grad Demo file!
-
If you are already familiar with PyTorch then you may be interested in the advanced notebook on Vision Foundation Models kindly provided by Li (Luis) Li; please see the instructions within the file, in particular using GPU acceleration when loading on the colab server.
-
There is also a folder (
SimpleDiffusion) with an example of a diffusion model kindly provided by Teo Deveney; this is currently setup to run on your local machine but you can take the contents of train.py and sample.py to run in collab if you want - there is also a checkpoint to restore the weights from if you just want to run the sampler. This code is based on the tutorial from Yang Song, referencing their paper, that you could also work through instead.