News for hackers. Stuff that matters.

A comment on Introducing Keras Core: Keras for TensorFlow, Jax, and PyTorch. Parent comment

Re:Introducing Keras Core: Keras for TensorFlow, Jax, and Py... by bootsmann on Tuesday July 11, 2023 @01:02PM (#36683867)
Pytorch is an animal by itself when you try to put it into production. They have started addressing it with torch 2.0 but it still has lengths to go. With this you can switch to TFserve if you have usual architecture.
[ Parent | on HN ]
Re:Introducing Keras Core: Keras for TensorFlow, Jax, and Py... by cygn on Tuesday July 11, 2023 @03:44PM (#36686298)
You can just use Triton which is basically TFserve for Tensorflow, Pytorch, Onnx and more.
[ Parent | on HN ]
Re:Introducing Keras Core: Keras for TensorFlow, Jax, and Py... by albertzeyer on Tuesday July 11, 2023 @04:27PM (#36686780)
Can you explain that?

My understand of Triton is more that this is an alternative to CUDA, but instead you write it directly in Python, and on a slightly higher-level, and it does a lot of optimizations automatically. So basically: Python -> Triton-IR -> LLVM-IR -> PTX.

https://openai.com/research/triton

[ Parent | on HN ]
Re:Introducing Keras Core: Keras for TensorFlow, Jax, and Py... by chillee on Tuesday July 11, 2023 @05:28PM (#36687455)
It's confusing, there's OpenAI Triton (what you're thinking of) and Nvidia Triton server (a different thing).
[ Parent | on HN ]
Re:Introducing Keras Core: Keras for TensorFlow, Jax, and Py... by jerrygenser on Tuesday July 11, 2023 @08:16PM (#36688884)
Original comment is referring to Nvidia triton inference server
[ Parent | on HN ]