Image to prompt with BLIP and CLIP
You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
|
|
|
|
import sys
|
|
|
|
|
from PIL import Image
|
|
|
|
|
from cog import BasePredictor, Input, Path
|
|
|
|
|
|
|
|
|
|
sys.path.extend(["src/clip", "src/blip"])
|
|
|
|
|
|
|
|
|
|
from clip_interrogator import Interrogator, Config
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
class Predictor(BasePredictor):
|
|
|
|
|
def setup(self):
|
|
|
|
|
config = Config(device="cuda:0", clip_model_name='ViT-L/14')
|
|
|
|
|
self.ci = Interrogator(config)
|
|
|
|
|
|
|
|
|
|
def predict(self, image: Path = Input(description="Input image")) -> str:
|
|
|
|
|
"""Run a single prediction on the model"""
|
|
|
|
|
image = Image.open(str(image)).convert("RGB")
|
|
|
|
|
return self.ci.interrogate(image)
|
