In this notebook, we will become familiar with:
- Tensor initialization
- Main attributes of tensors: dtype, device, requires_grad
- Tensor operations
- Derivatives and gradients
- Simple linear regression
- Train and Fit
In this notebook we will solve the digit classification task with 3 different models of progressively increasing complexity:
- Linear classifier (logistic regression)
- Multilayer perceptron (non-linear fully connected network)
- Convolutional neural network
In this notebook we will try to build real world classifier using Convolutional neural network.
Links: