A Repo Containing Simple Practices To Start On Deep Learning

• training and analysis of deep neural networks using several major dl models for cv and nlp as examples (mostly in PyTorch, but also in TensorFlow)
• build a simple conv-net, optimizer, etc., in Numpy and train it
• build several custom-defined modules not included as built-in methods in PyTorch by extension on its autograd system
• homework / exam solutions with codes for several online courses on reinforcement learning
• train a deep reinforcement learning agent in OpenAI's Gym platform
• build implementations for several common RL algorithms, following major RL lib's
• implement and examine the ENAS framework in PyTorch following the original paper