feat: 🚀 from_vlm deepseek vl 2 support

[onuralpszr]

Description

• Deepseek VL2 support added by @SkalskiP
• Deepseek VL2 test added @onuralpszr

Type of change

• New feature (non-breaking change which adds functionality)
• This change requires a documentation update

How has this change been tested, please provide a testcase or example of how you tested the change?

The code changes in this PR was tested primarily with the following colab:

The code can also be tested with the following simple script with a result formatted like DeepSeek-VL2's response:

# wget https://huggingface.co/spaces/deepseek-ai/deepseek-vl2-small/resolve/main/images/visual_grounding_1.jpeg

from PIL import Image
import supervision as sv


image = Image.open("visual_grounding_1.jpeg")
result = "<|ref|>The giraffe at the back<|/ref|><|det|>[[580, 270, 999, 904]]<|/det|><|ref|>The giraffe at the front<|/ref|><|det|>[[26, 31, 632, 998]]<|/det|><｜end▁of▁sentence｜>"
dets = sv.Detections.from_vlm(vlm=sv.VLM.DEEPSEEK_VL_2, result=result, resolution_wh=image.size)
box_annotator = sv.BoxAnnotator()
label_annotator = sv.LabelAnnotator(smart_position=True)
annotated_image = box_annotator.annotate(scene=image, detections=dets)
annotated_image = label_annotator.annotate(scene=annotated_image, detections=dets)
annotated_image.save("annotated_image.png")

print(dets.xyxy)
print(dets.class_id)
print(dets.data)

Docs

• Docs updated