12 lines
303 B
Python
12 lines
303 B
Python
from moondream import VisionEncoder, TextModel
|
|
from PIL import Image
|
|
|
|
vision_encoder = VisionEncoder()
|
|
text_model = TextModel()
|
|
|
|
image = Image.open("assets/demo-1.jpg")
|
|
image_embeds = vision_encoder(image)
|
|
out = text_model.generate(image_embeds, "User: <image>\nWhat is this?\nAssistant: ")
|
|
|
|
print(out)
|