sdaxixihaha

Hi,image_features = model.encode_image(image) and the output of encode_image is a tensor of dimension 512. Do you know how to control dimension of output? eg, I want 128 or 64 dimension...