4 ms·

I’ve also ran models in Go, transformers even T5. There wasn’t that much overhead maybe some annoying compilation stuff but nothing crazy. This was tensorflow

by ramoz 3y ago

I’ve also ran models in Go, transformers even T5. There wasn’t that much overhead maybe some annoying compilation stuff but nothing crazy. This was tensorflow btw which has Go bindings support. It is a smart & worthwhile move, we also needed to drop python for performance/cost gains.

astronautas 3y ago
eh, awesome! Seems this one, right? https://github.com/galeone/tfgo https://github.com/galeone/tfgo. Quite many stars.
- ramoz 3y ago
  I think just native https://pkg.go.dev/github.com/tensorflow/tensorflow/tensorflow/go https://pkg.go.dev/github.com/tensorflow/tensorflow/tensorfl... but tfgo looks interesting. Actually the docs around this weren’t great. Took the train-in-python & inference-in-go approach. And only for versions greater than tf2
  - astronautas 3y ago
    Write a blog post then about this! I can tell you it is hardly a solved problem.