Hi, I’ve been running into issues converting my .onnx model into a Hef model. The main cause is the nvidia gpu requirement. At least that’s what i’ve gathered. I have tried with the docker, with all versions of python, and all versions of the software wheels available for download. However I’m still not able to convert it. Would it be possible for someone to create a google collab for this conversion?
many thanks
Hey @user149,
Can you try to parse the model? This step doesn’t require a GPU. If you already have the ONNX converted to HAR (parsed), you can send it to me and I can try to compile it on my end.
You can run the DFC parsing step without an NVIDIA GPU. It’s specifically the optimization step that requires the GPU. The compilation step can also be run without it, so if you provide the parsed file, I can run the optimization step for you.