Fine-Grained Visual Classification on FGVC-Aircraft

Project Page: SelfSynthX.

Paper on arXiv: Enhancing Cognition and Explainability of Multimodal Foundation Models with Self-Synthesized Data

This model is a fine-tuned multimodal foundation model based on LLaVA-1.5-7B-hf, optimized for fine-grained classification of aircraft types using the FGVC-Aircraft dataset.

Key Details

Base Model: LLaVA-1.5-7B
Dataset: FGVC-Aircraft (Fine-Grained Visual Classification of Aircraft)
Innovation:
- Self-Synthesized Data: Extracts and highlights distinctive aircraft-specific visual features using the Information Bottleneck principle.
- Iterative Fine-Tuning: Uses reward model-free rejection sampling to improve classification accuracy and explanation quality.
Intended Use: Identification of aircraft models with human-verifiable explanations.

How to Use

import requests
from PIL import Image
import torch
from transformers import AutoProcessor, LlavaForConditionalGeneration

model_id = "YuchengShi/LLaVA-v1.5-7B-Fgvc"
model = LlavaForConditionalGeneration.from_pretrained(
    model_id,
    torch_dtype=torch.float16,
    low_cpu_mem_usage=True,
).to("cuda")
processor = AutoProcessor.from_pretrained(model_id)

conversation = [
    {
      "role": "user",
      "content": [
          {"type": "text", "text": "What type of aircraft is this?"},
          {"type": "image"},
        ],
    },
]
prompt = processor.apply_chat_template(conversation, add_generation_prompt=True)
image_file = "fgvc-aircraft/test1.png"
raw_image = Image.open(requests.get(image_file, stream=True).raw)
inputs = processor(images=raw_image, text=prompt, return_tensors='pt').to("cuda", torch.float16)

output = model.generate(**inputs, max_new_tokens=200, do_sample=False)
print(processor.decode(output[0][2:], skip_special_tokens=True))

Training & Evaluation

Training: Fine-tuned using LoRA on FGVC-Aircraft with iterative rejection sampling.
Evaluation: Achieves high accuracy in distinguishing aircraft types while providing detailed, interpretable explanations.

Citation

If you use this model, please cite:

@inproceedings{
  shi2025enhancing,
  title={Enhancing Cognition and Explainability of Multimodal Foundation Models with Self-Synthesized Data},
  author={Yucheng Shi and Quanzheng Li and Jin Sun and Xiang Li and Ninghao Liu},
  booktitle={The Thirteenth International Conference on Learning Representations},
  year={2025},
  url={https://openreview.net/forum?id=lHbLpwbEyt}
}