Dgv2 commited on
Commit
6f7a0df
·
verified ·
1 Parent(s): 7c46844

Create Project_README.txt

Browse files
Files changed (1) hide show
  1. Project_README.txt +23 -0
Project_README.txt ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Dog Breed Identifier 🐶
2
+
3
+ ## Live App
4
+ Try it here: https://huggingface.co/spaces/Dgv2/Final_Project
5
+
6
+ ## Project Description
7
+
8
+ I decided to do this project as my final assignment because of my love for animals, especially dogs. One of the earlier homework assignments involved video recognition, and I actually used footage of my own dog — it was one of my favorite parts of the course. So when I had the chance to build a final project around something I care about, this felt like the perfect fit. I really appreciate being able to wrap up the course with something that combines what I’ve learned with what I enjoy.
9
+
10
+ The app I built is a simple tool where users can upload a photo of a dog, and it will predict the breed from a set of 120 possibilities. It uses a Vision Transformer model (ViT) that was pre-trained and fine-tuned specifically for dog breed classification. I put everything together using Gradio for the interface and deployed it on Hugging Face Spaces to make it accessible online. Overall, this project helped me apply image classification techniques from class in a hands-on way, while getting to work on a topic that’s genuinely fun and interesting to me.
11
+
12
+ ## How It Works
13
+
14
+ - Upload an image of a dog.
15
+ - The app predicts one of 120 dog breeds using a pretrained ViT model.
16
+ - Built using Python, Transformers, Torch, and Gradio.
17
+
18
+ ## Tools Used
19
+
20
+ - Hugging Face Spaces
21
+ - Gradio
22
+ - Vision Transformer (ViT)
23
+ - Stanford Dogs dataset