Documentation support for Crestereo Pytorch model conversion to onnx model

Issue Type

Support

OS

Ubuntu

OS architecture

x86_64

Programming Language

Python

Framework

PyTorch, ONNX

Model name and Weights/Checkpoints URL

https://github.com/PINTO0309/PINTO_model_zoo/tree/main/284_CREStereo

Description

Hi, Thank you for your great work.

Can you guide me through the procedure for converting CreStereo PyTorch model to Onnx model.? As per the documents, it says You need to convert the model twice, once without the init_flow and once with the init_flow. So the commented out part in your script are not options, but two separate runs. Then use the merging tool PINTO0309 developed to merge the two parts into one. Or, just use them separately in sequence.

But in the second step while using onnx_merge.py ,

i am facing the issue attached in the below error log

Relevant Log Output

Traceback (most recent call last):
  File "onnx_merge.py", line 17, in <module>
    combined_model = onnx.compose.merge_models(
  File "/home/virtual_environments/cre_pytorch/lib/python3.8/site-packages/onnx/compose.py", line 363, in merge_models
    graph = merge_graphs(
  File "/home/virtual_environments/cre_pytorch/lib/python3.8/site-packages/onnx/compose.py", line 187, in merge_graphs
    raise ValueError(f"Input {g2_in_name} is not present in g2")
ValueError: Input next_flow_init is not present in g2

URL or source code for simple inference testing code

No response

PINTO0309 / PINTO_model_zoo