[Bug] image_tokens not reported when using multi modal calls

[delip]

What happened?

Steps to reproduce

image_captioner = dspy.Predict(dspy.Signature("image:dspy.Image -> caption:str", instructions="Generate a caption for an image"))
with dspy.context(track_usage=True):
    response = image_captioner(image=dspy.Image.from_url("https://picsum.photos/id/45/200/300"))
    print(response)
    print(response.get_lm_usage())

DSPy version

3.0.0

[TomeHirata]

DSPy encodes the image into a text when calling LLM. So it is expected that text token is used instead of image token.