Skip to content

Latest commit

 

History

History
25 lines (19 loc) · 1.76 KB

README.md

File metadata and controls

25 lines (19 loc) · 1.76 KB

Pet Breed Classifier

This demo is hosted here.

Steps to recreate

  1. create a new conda environment: conda create --name myenv
  2. activate the new env: conda activate myenv
  3. install all the dependencies pip install -r requirements.txt
  4. create your own model on Google Colab with train.py and place the .pkl file in the working directory
  5. run the app.py file: python app.py

Training your model

You can pick an option that's convenient to you based on the system you are using. If you're using a not so powerful system, you may want to use Colab or Kaggle for GPUs to prevent spending time debugging avoidable/out-of-memory errors.

  1. sample code from Fast AI course: https://shorturl.at/lyDR8
  2. another version can be referred here
  3. use the train.py code from this repo in Google Colab

Once you have trained the model, place the pkl file in the working directory.

Limitations

This code is compatible with UNIX based systems. If you're running a windows machine and you're getting a NotImplementedError: cannot instantiate 'PosixPath' on your system, then you may want to refer the another version of the code that's compatible with windows machines. The repo can be downloaded here. Note- This file is ~800 MB and it has a .pth file instead of .pkl file

More

  1. Which image models are best? | Kaggle
  2. The best vision models for fine-tuning | Kaggle