Skip to content

Latest commit

 

History

History
29 lines (26 loc) · 1.87 KB

getting_started.md

File metadata and controls

29 lines (26 loc) · 1.87 KB

Installation :

Tip : Please make sure to install all dependencies from requirements.txt.
Grab a cup of coffee as these will take some time !

Get, set and go :

  • Download complete Project files using following command from git bash:
git clone https://github.com/souravs17031999/Dog-Breed-Classifier-App
  • Open terminal (cmd) and then move to Project root directory.
  • Now, run the following file "dog_breed_predict.py" using following command (which also gives path of image supplied) :
python dog_breed_predict.py <PATH>

where python (python3) depending upon env and is simply the image actual path you supply (without "<", ">").
Ex.

python dog_breed_predict.py dog.JPG
  • There you go, predicted label name along with the image you supplied will be printed on python console.

For live predictions :

  • IMPORTANT : If you want to run live predictions from feed taken from your live cam, then instead of file "dog_breed_predict.py" which gives predictions on input images, you should run "dog_breed_predict_cam.py"

NOTE : you should already have weights of pre trained model used in this project (mail me at [email protected] for model_transfer.pt file) and replace the following line with your model_transfer.pt file path and also you need "resnet50.pth" file which is already available here at kaggle (download this).

model_transfer = load_model('model_transfer.pt')

TIP : Keep everything in the same project folder and enjoy !