-
Notifications
You must be signed in to change notification settings - Fork 0
/
masks_to_text.py
42 lines (35 loc) · 1.36 KB
/
masks_to_text.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
from transformers import BlipProcessor, BlipForConditionalGeneration
from PIL import Image
import torch
class BlipImageCaptioning:
def __init__(self, model_path="models/BLIP"):
self.device = "cpu"
if torch.cuda.is_available():
self.device = "cuda"
elif torch.backends.mps.is_available():
self.device = "mps"
# Initialize the processor and model
self.processor = BlipProcessor.from_pretrained(model_path)
self.model = BlipForConditionalGeneration.from_pretrained(model_path).to(
self.device
)
def load_image(self, image_path):
"""
Loads an image and converts it to RGB.
"""
return Image.open(image_path).convert("RGB")
def generate_description(self, image, max_new_tokens=512):
"""
Processes the image and generates a description.
"""
inputs = (
self.processor(image, return_tensors="pt").to(torch.int32).to(self.device)
)
output = self.model.generate(**inputs, max_new_tokens=max_new_tokens)
description = self.processor.decode(output[0], skip_special_tokens=True)
return description
if __name__ == "__main__":
blip = BlipImageCaptioning()
img = blip.load_image("images/telephone_booth.jpg")
description = blip.generate_description(img)
print(description)