Hi,
I want to load the pose estimation 1b model for 133 key points independently from the .sh script. But it need the class of model's architecture because it is saved in .pth format. But for the 308 key points, because it is saved in .pt format, I easily load the model and run it for my inference without running the .sh script. My question is why you did not save the 133 key points in .pt format and how can I find the model's architecture class in case if there is not any .pt format for the saved model.
Hi, I want to load the pose estimation 1b model for 133 key points independently from the .sh script. But it need the class of model's architecture because it is saved in .pth format. But for the 308 key points, because it is saved in .pt format, I easily load the model and run it for my inference without running the .sh script. My question is why you did not save the 133 key points in .pt format and how can I find the model's architecture class in case if there is not any .pt format for the saved model.