调用Blip image tokenizer实现image2text

登录后复制

from transformers import BlipProcessor, BlipForConditionalGeneration
from PIL import Image

# Load the model and processor
processor = BlipProcessor.from_pretrained("huggingface.co/Salesforce/blip-image-captioning-base")
model = BlipForConditionalGeneration.from_pretrained("huggingface.co/Salesforce/blip-image-captioning-base")

# Load and preprocess an image
img = Image.open("data/input_image.png")
inputs = processor(img, return_tensors="pt")

# Generate caption
out = model.generate(**inputs)
caption = processor.decode(out[0], skip_special_tokens=True)

# Print the generated caption
print(caption)

• 1.
• 2.
• 3.
• 4.
• 5.
• 6.
• 7.
• 8.
• 9.
• 10.
• 11.
• 12.
• 13.
• 14.
• 15.
• 16.
• 17.

原创作者: guotong1988 转载于: https://blog.51cto.com/guotong1988/11830188