This is more of a question than an issue. I want to train the model (Owl-ViT) on custom dataset from checkpoint ("google/owlvit-base-patch32" or clip_vit) and want to use it in custom pipeline. For that I want to convert this model into ".engine" file. Is there any way of saving the model in a particular format and convert it to ONNX format?
Hi all,
This is more of a question than an issue. I want to train the model (Owl-ViT) on custom dataset from checkpoint ("google/owlvit-base-patch32" or clip_vit) and want to use it in custom pipeline. For that I want to convert this model into ".engine" file. Is there any way of saving the model in a particular format and convert it to ONNX format?
Thank you.
Best regards, Akash.