Skip to content
This repository has been archived by the owner on May 6, 2022. It is now read-only.

Add PyTorch model abstraction #100

Merged
merged 1 commit into from Jun 30, 2021
Merged

Add PyTorch model abstraction #100

merged 1 commit into from Jun 30, 2021

Conversation

will-rice
Copy link
Member

@will-rice will-rice commented Jun 30, 2021

PyTorch is a popular ML framework in Python. Users may want to be able to use jit traced PyTorch models with Spokestack. This abstraction offers users that ability.

@will-rice will-rice changed the title Add Pyorch model abstraction Add PyTorch model abstraction Jun 30, 2021
Copy link
Contributor

@space-pope space-pope left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This allows PyTorch models to be loaded, but doesn't do much with them. The readme instructions explain how to load a model, but not how to use one in the context of Spokestack. Are further integrations/instructions planned for future PRs?

In other words, this is more of a library contributor/developer feature as written, something that you'd mention in tutorials about building your own integrations rather than the readme.

@will-rice
Copy link
Member Author

This is intended to be the complement to the TFLite model abstraction. Ultimately, I think we could let the user choose between TensorFlow or PyTorch. I will take out the documentation in the README, because like you said only a contributor really needs to know about this class.

@space-pope space-pope merged commit 0491348 into master Jun 30, 2021
@space-pope space-pope deleted the wr-torch branch June 30, 2021 17:13
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants