SmolVLM2-500M FoodExtract LoRA
LoRA adapter for HuggingFaceTB/SmolVLM2-2.2B-Instruct,
fine-tuned on mrdbourke/FoodExtract-1k-Vision.
Given a food/drink image it returns structured JSON:
{"is_food": 1, "image_title": "salad",
"food_items": ["tomato", "lettuce"], "drink_items": []}
Usage
from transformers import AutoProcessor, AutoModelForImageTextToText
from peft import PeftModel
base = "HuggingFaceTB/SmolVLM2-2.2B-Instruct"
processor = AutoProcessor.from_pretrained(base)
model = AutoModelForImageTextToText.from_pretrained(base)
model = PeftModel.from_pretrained(model, "sergiudanstan/smolvlm2-2.2b-foodextract-lora").merge_and_unload()
- Downloads last month
- 6
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for sergiudanstan/smolvlm2-2.2b-foodextract-lora
Base model
HuggingFaceTB/SmolLM2-1.7B Quantized
HuggingFaceTB/SmolLM2-1.7B-Instruct Quantized
HuggingFaceTB/SmolVLM-Instruct Finetuned
HuggingFaceTB/SmolVLM2-2.2B-Instruct