concurrent_inference

An example of how to use the multiprocessing package along with PyTorch.

Code pertaining to this Medium post.

---

What does it do?

• A processes [R] reads images from a folder [Fo] and multiple detection processes [D#] are used to obtain the class wise count of objects in the images and write it to a file [Fi].
• It uses torch.multiprocessing for multiprocessing, PIL.Image to read the images and tqdm to keep track of the queue.
  

Usage

• Basic usage : $ python count_objects.py -f input_folder -o output_file.log
• For other options : $ python count_objects.py -h