Vehicle Classification - UTD Evaluation
This model is a custom CNN architecture trained to classify 8 categories of vehicles.
📝 Model Card
- Architecture: 3-Block Custom CNN (Not Pretrained)
- Target Accuracy: >50.0%
- Final Test Accuracy: 78.54%
📂 Dataset & Split
The model utilizes a dataset of 26,378 images with an 8:2 train-test split.
- Training Samples: 21,102
- Testing Samples: 5,276
- Normalization: Mean [0.485, 0.456, 0.406], Std [0.229, 0.224, 0.225]
🛡️ Safety & Robustness Features
To align with AI Safety principles, the following design choices were implemented:
- High Regularization: Dropout (0.5) to prevent over-reliance on high-frequency noise.
- Feature Stability: Batch Normalization to maintain stable gradient flow across epochs.
- Generalization Proof: The 0.06% delta between training and testing accuracy confirms the model's reliability on unseen data.
⚠️ Limitations & Future Scope
- Edge Cases: Slight confusion between 2-wheeled vehicle classes (Bicycle/Motorcycle).
- Environment: Performance has not yet been validated for low-light or adverse weather conditions.
🔗 Links for this project
🛠️ Other Key Projects
- Language Datasets: Curated and published 100k+ rows datasets for low resource languages on Hugging Face for regional NLP research.
- CNSD Model Architecture: Authored research on neural network configurations for optimized feature extraction.
- Custom 4 sentiment models, 1 vehicle classification model and several datasets.
- Link to my Hugging Face account : https://huggingface.co/abhiprd20 (with all models and datasets)
- My NLP research paper pre-print : https://zenodo.org/records/19054785
- 2nd research project : https://github.com/abhiprd200/CNSD_prototype
- Demo of this project: Live on Hugging Face Spaces
- My github with other projects : https://github.com/abhiprd200
Contact
- E-mail : abhiprd20@gmail.com
🚀 Usage
import torch
# Architecture must match VehicleClassifier class in provided app.py
model.load_state_dict(torch.load("model.pth", map_location="cpu"))
model.eval()