Hello everyone! I would like to convert the CLIP model to ONNX format. I read in the documentation how to do it, and this is what happened: My libs: torch - 1.12.1 transformers - 4.23.1 onnxruntime - 1.11.1 onnxsimplifier - 0.4.8 Code: import time import PIL import torch import onnx import on…

you can try: inputs = processor(text=["a photo of two cat"], images=image, return_tensors="np", padding=True) session.run(input_feed=dict(inputs), output_names=["logits_per_image"]) which returns logits_per_image for each of the text inputs

Troubles with converting CLIP to ONNX

Beginners

mineshj1291 October 21, 2022, 11:15am 4

are you looking for this:

!python -m transformers.onnx --model=local-pt-checkpoint onnx/

Topic		Replies	Views
Convert model clip to onnx Models	0	214	July 5, 2024
How to give CLIP only image as input? (without text) Beginners	0	584	October 25, 2022
How to successfully ONNX pretrained models Beginners	7	315	January 26, 2025
Converting SWIN Transformers from Pytorch through ONNX or Others 🤗Transformers	0	490	May 16, 2023
Is there a way to include the text_projection and/or embedding normalization in an optimum-optimized CLIPTextModelWithProjection? 🤗Optimum	8	1290	July 1, 2023

Troubles with converting CLIP to ONNX

Related topics