AI Saturdays Uyo: Persistence, Patience and Perspiration Make Unbeatable Combination For Success. Napoleon Hill.
We had our second meetup on 17th February 2017. People came in numbers 1,2,3,4…………5, although, not all that signed up turn-up. At this point we did not wait for crowd again because “time wait for no man”.
We started by 10:00 am and ended by 4:00 pm all the session was taken. 10:00 am to 11:00 am was used to quickly review last meetup lesson and welcome new participants who joined us.
Fast.ai (practical deep learning for coders lesson 2 was taken) from 11:00am-12:30pm. where we get to know more about the model, learning rate, epoch, loss validation, how to fine tune a network and lastly how to setup aws (amazon web service) thanks to Jeremy Howard.
We quickly moved to convolutional neural network for visual recognition lecture one where we learnt about what the course is all about that is, computer version: the study of visual data. We also learnt the history of computer version and general overview of the course. Fill free to check out the video here. This session lasted for one hour thirty minute .from 12:30pm-2:00pm.
Thirty minute was use for break. During this thirty minute we decided to put some participants who where not familiar through on Git. A big thanks to Onuwa Nnachi Isaac.
Session three, Reinforcement learning by David Silver lecture one which appeared to be the most interesting session to majority of the participant . The video briefly introduce what a Reinforcement learning is, problems within reinforcement learning agent, the reinforcement learning problems. It also cover some reinforcement learning key words like environment, agent, reward , action and observation.
We are staying with three session for now and may introduce some letter depending on the participant interest. I will be very self-fish if I fall to appreciate nature.ai for this beautiful opportunity and my fellow ambassador for their time. Wils Wilson and Jude Ben.
Reference
Ultimate guide to setting up a Google Cloud machine for fast.ai version 2