README.md

metadata

title: LN2 Computer Vision
emoji: 💻
colorFrom: pink
colorTo: indigo
sdk: gradio
sdk_version: 5.25.2
app_file: app.py
pinned: false

🐶 Zero-Shot Classification on Oxford-IIIT Pet Dataset

This model was evaluated on the Oxford-IIIT Pet Dataset in a zero-shot setting using Hugging Face's 🤗 transformers and datasets libraries.

🧪 Task: Zero-Shot Image Classification

In this task, the model was used without any fine-tuning to classify pet images into breed categories. Each class label was passed as a candidate for classification using natural language descriptions. The model selected the most likely label for each image based on its learned knowledge.

✅ Results

Metric	Value
Accuracy	88.00%
Precision	87.68%
Recall	88.00%

⚠️ Note: These results are based on zero-shot inference and do not reflect fine-tuned performance.

🧠 Model Used

We used the pre-trained openai/clip-vit-base-patch32 for zero-shot classification.