Convolutional neural nets are the next thing in the deep learning world. They are the reason why deep learning is so popular in the field of computer vision. They are the reason why we can do so many things with images.

Things to look for while building deep learning models

  1. Gather the dataset.
  2. Clean the dataset.
  3. Split the dataset into train, test and validation.
  4. Build the model.
  5. Train the model.
  6. Evaluate the model.
  7. Deploy the model.
  8. Monitor the model.
  9. Improve the model.
  10. Repeat the process. Process of building a CNN

Getting started with the convolutional tasks

  1. Math for convolution - Link
  2. Pytorch to implement a basic CNN - Link
  3. Hyperparameters
    1. Optimizers - Math 5.x series
    2. Regularization - Math 8.x, 9.x series
    3. Augmentations - Augmentations
  4. Model Experimentations
    1. wandb - Link
    2. comet - Link
  5. Model Interpretability - Link
  6. Deploy the model - Link
  7. Monitor the model - Link

Image Processing Blogs

  1. Chapter 1
  2. Chapter 2
  3. Chapter 3

Deep Learning Talks

  1. Using Images to train LLM models.
  2. Transfer Learning from nature.
  3. Underwater robot navigation.
  4. Model Training with zero data.