How to actually use this in pytorch?
I'm relatively new to AI, but have trained an LSTM network myself in PyTorch.
But I'm not sure how to use this model in a programmatic way (I can't use a webui since I have tens of thousands of images I need to generate, so I have no intentions of doing this by hand).
Using the safetensors module I can safe_open the model, but not deserialize it (complains about missing metadata), and safe_open doesn't give you the ability to do inferences, all it lets you do is pick individual tensors and see the hundreds of tensor keys, and it's not clear to me what tensors I would like to pick/pickle either.
Edit: Figured out how to use it in PT following this tutorial https://huggingface.co/docs/diffusers/main/en/using-diffusers/img2img... still not sure how to retrain the model, but I guess I'll figure it out.
Hi, the only way I know to use models is to uploading them here, transforming them to diffusers and using their inference API, so I'm unable to help beyond this point, but I hope you can find a way to do what you want.