How to upgrade torchvision to newest unstable version (0.12)?

Hello, I was happy to find pytorch dataset support for the Describable Texture Dataset (DTD) on the docs.

However, after I’ve failed trying to import it in my google colab instance (using torch v1.10 and torchvision 0.11) I noticed that it was only available on the main (0.12) unstable version of torchvision.

So what should I do to use it?

When I try:

!pip install --upgrade torchvision==0.12

I get:

ERROR: Could not find a version that satisfies the requirement torchvision==0.12 (from versions: 0.1.6, 0.1.7, 0.1.8, 0.1.9, 0.2.0, 0.2.1, 0.2.2, 0.2.2.post2, 0.2.2.post3, 0.3.0, 0.4.0, 0.4.1, 0.4.2, 0.5.0, 0.6.0, 0.6.1, 0.7.0, 0.8.0, 0.8.1, 0.8.2, 0.9.0, 0.9.1, 0.10.0, 0.10.1, 0.11.0, 0.11.1, 0.11.2, 0.11.3)
ERROR: No matching distribution found for torchvision==0.12

Since it is the unstable version, we would have to build it

Alternatively, you could also install the nightly binaries instead of building from source.

1 Like