DL with PyTorch (Week 2)

Done with Assignment 2 🤩🎉😎🎊

Although I am running a bit late but still I will try to complete the assignments and respective blogs ASAP.

This assignment was a part of the 6 week-long free course named, Deep Learning with PyTorch: Zero to GANs provided by Jovian.ml in collaboration with FreeCodeCamp.

Learned about linear regression, loss functions and we trained a model using linear regression to do insurance cost prediction.

My key observations🧐 during this assignment were:

• It's better to choose loss function from below list if you are working on linear regression
  • torch.nn.functional.l1_loss
  • torch.nn.functional.mse_loss
  • torch.nn.functional.smooth_l1_loss
• Larger the value of no. of epochs and learning rate , the more rapid changes can be seen in the loss_value

You can find a detailed explanation here Have a look at it.
To find out what I have been doing in this Lock-down period, click here