CPU support by ONNX as phi3?

by alierenak - opened Aug 21

Aug 21

As far as I know, this model currently doesn't support running on a CPU due to flash_attn constraints. Are there any plans to release an ONNX version, similar to what was done with phi3

LeeStott

Aug 21

3.5 versions will ship too https://huggingface.co/onnx-community

At present there is a 3.5 mini instruct onnx model available https://huggingface.co/onnx-community/Phi-3.5-mini-instruct-web/tree/main

phi0112358

Aug 21

the second link leads to 404 ('-onnx' ist missing)

this link does work:
https://huggingface.co/onnx-community/Phi-3.5-mini-instruct-onnx-web/tree/main

Srrp

Aug 23

How to run onnx model

nguyenbh changed discussion status to closed Oct 10

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment