README.md

metadata

license: apache-2.0
tags:
  - image-captioning
  - furniture
  - blip
  - vision-language
model-index:
  - name: furniture-captioner
    results: []

furniture-captioner

Model Description
furniture-captioner is a fine-tuned version of BLIP specialized in generating captions for furniture images.
It has been trained using the dataset yemalin/furniture_ds.

Use Cases

• Generating product descriptions for furniture marketplace listings
• Improving searchability through auto-generated captions
• Enhancing accessibility with alternative text

Training
Fine-tuned starting from a pre-trained BLIP model on a curated furniture dataset.
All images are annotated with relevant captions (design, style, function).

Intended Uses & Limitations

• ⚡ Works best on images of indoor and outdoor furniture.
• 🚫 Not optimized for general objects or human activities.

License
Apache 2.0 — allowing commercial and non-commercial usage with attribution.

---

Usage Example

from transformers import BlipProcessor, BlipForConditionalGeneration
from PIL import Image
import requests

processor = BlipProcessor.from_pretrained("yemalin/furniture-captioner")
model = BlipForConditionalGeneration.from_pretrained("yemalin/furniture-captioner")

img_url = "https://example.com/your-furniture-image.jpg"
image = Image.open(requests.get(img_url, stream=True).raw)

inputs = processor(images=image, return_tensors="pt")
out = model.generate(**inputs)
caption = processor.decode(out[0], skip_special_tokens=True)
print(caption)