DogBreedSpotter is a Python-based image classification project designed to identify and classify dog breeds in images. This project utilizes deep learning models, including convolutional neural networks (CNNs) such as VGG, AlexNet, and ResNet, to accurately detect whether an image contains a dog and, if so, determine the breed.
- Image classification to identify dogs
- Accurate breed classification for dog images
- Utilizes pre-trained deep learning models
- Measures algorithm accuracy and runtime performance
- Supports various dog breeds recognition
The primary goal of DogBreedSpotter is to provide a reliable tool for dog breed identification, making it useful for various applications, including dog shows, pet registration, and more.
- Clone the repository.
- Install the required dependencies.
- Run the project and provide an image for breed identification.
- Explore the results and accuracy of different deep learning models.
In order to deploy your Flask API to the cloud, here are the steps that you need to follow:
Obviously, the first step is to create your web application.
Then, when your Flask app is implemented, you can launch it locally using the following command:
flask run
This is a development server, you cannot use it in a production deployment. In production
Boot disk size should be at least 30GB
- Host: Your VM name (from GCP Console)
- HostName: Your VM External IP Address (from GCP Console)
- User: The username you want to use to SSH to this VM
- IdentityFile: The path to your IdentityFile
mkdir google_compute_engine
cd google_compute_engine
ssh-keygen -t rsa -f remote-ssh-demo -C omarkhalil -b 2048
cd google_compute_engine
cat remote-ssh-demo.pub
- Copy the content of [remote-ssh-demo.pub]
- Open GCP Console
- Go to Compute Engine Instances
- Click on your instance (in our case classifier-vm)
- Click on Edit
- Scroll down to SSH Keys
- Click on ADD ITEM
- Paste the content you copied from remote-ssh-demo.pub into SSH Key 1
- Click [Ctrl+Shift+p]
- Click on Connect to Host
- Choose classifier-vm
git clone https://github.com/OmarKhalil10/DogBreedSpotter.git
It is always a good idea to update before trying to install a new package. Run the commands below:
sudo apt-get update
Upgrade the dependencies using the command:
sudo apt upgrade
If Python 3 has already been installed on the system, execute the command below to install pip3:
sudo apt-get -y install python3-pip
Verification (Verify your installation)
To verify the installation, run the following command to cross check the version number:
pip3 --version
pip install --upgrade pip virtualenv
Then make virtualvenv
python -m virtualenv venv
Then you will need to activate the virtaual env
source venv/bin/activate
After that you should install dependencies using the command
pip install -r requirements.txt
set FLASK_APP='app.py'
flask run --reload
Ensure that it works correctly and close it
export [email protected]
export MY_PROJECT_ID=landmark-classifier
gcloud config set account $MY_EMAIL_ADDRESS
gcloud auth login $MY_EMAIL_ADDRESS
gcloud config set project $MY_PROJECT_ID
Create the Dockerfile
# Use Python37
FROM python:3.7
## Step 1:
# Create a working directory
WORKDIR /app
## Step 2:
# Copy source code to working directory
COPY app.py requirements.txt /app/
COPY templates /app/templates/
COPY static /app/static/
COPY uploads /app/uploads/
COPY ./imagenet-dictionary /app/imagenet-dictionary/
## Step 3:
# Install packages from requirements.txt
RUN pip install -r requirements.txt && \
rm /app/requirements.txt
## Step 4:
# Expose port 8080
EXPOSE 8080
# Run app.py at container launch
CMD ["python", "app.py"]
sudo -s
Run the following command to install docker
sudo apt-get install docker.io
When Docker is ready! Test docker
docker ps
We shall build the Docker image and push it to Container Registry on GCP. The service name is defined as dogbreedspotter with tag v1
docker login
sudo bash run_docker.sh
If you face a problem with the size of the disk no space left Check Disk Usage:
Use the df -h command to check the current disk usage on your instance. This command will show you the disk space usage for each mounted filesystem.
df -h
Identify the filesystem with limited space, and take appropriate action to free up space on that filesystem.
If you've determined that you need more disk space for your Compute Engine instance, you can resize the root disk. Here are the general steps:
sudo poweroff
If you are logged out you could srart the instance again from the console after increasing the disk space
- In the Google Cloud Console, navigate to "Compute Engine" > "Disks."
- Select the disk attached to your instance.
- Click "Edit" and increase the size of the disk as needed.
- Start the Compute Engine instance:
If You want to start the Compute Engine instance from your terminal
sudo poweron
docker build -t gcr.io/$MY_PROJECT_ID/dogbreedspotter:v6 -f Dockerfile .
docker push gcr.io/$MY_PROJECT_ID/dogbreedspotter:v6
gcloud run deploy dog-breed-spotter \
--image gcr.io/$MY_PROJECT_ID/dogbreedspotter:v6 \
--region europe-west6 \
--platform managed \
--memory 8Gi \
--cpu 2 \
--max-instances 25
Use the gcloud run revisions list command to list all revisions of your Cloud Run service. Replace <SERVICE_NAME> with the name of your service.
gcloud run revisions list --platform managed --region europe-west6 --service dog-breed-spotter --format="value(name)" | sort
Use the gcloud run revisions delete command to delete each of the old revisions. Replace <REVISION_NAME> with the name of each revision you want to delete. [If Any!]
gcloud run revisions delete <REVISION_NAME> --platform managed --region europe-west6 --quiet
You can run this command for each old revision you copied in the previos step
gcloud run revisions describe dog-breed-spotter-00019-5s2 \
--platform managed \
--region europe-west6 \
https://dog-breed-spotter-t6mbhrffxa-oa.a.run.app
Or you send HTTPS requests to the Cloud Run instance for testing
python request_main_v6.py
Contributions and enhancements to DogBreedSpotter are welcome! Feel free to fork the repository, make improvements, and submit pull requests.
This project is licensed under the MIT License. See the LICENSE file for details.
- [Omar Khalil]
If you have any questions or suggestions, please feel free to contact me.