`import`: Support image as input type.

[daavoo]

For computer vision models, it would be very handy to support image paths as input:

mlem apply visionmodel image.png -i --it image[tensorflow] -o prediction

Both TensorFlow and Pytorch provide utils for loading image paths into Tensors:

• https://www.tensorflow.org/api_docs/python/tf/keras/utils/load_img#usage
• https://pytorch.org/vision/stable/generated/torchvision.io.read_image.html#torchvision.io.read_image

[aguschin]

Thanks for mentioning this @daavoo! @mike0sv, I think this should be addressed somewhere in Q3.

[mike0sv]

related https://github.com/iterative/mlem/issues/332