In checkpoints model video

if we are training for more than 1 epoch than where do we apply checkpoints
after every iteration or after every epoch

If you are using PyTorch, you can place the checkpoints after this optimizer.step() statement. In general you should place checkpoints just after updating the parameters.