I never had the idea that one day I would start my AI and deep learning journey with fast.ai, actually, I never knew I would even learn it. However, I’ve already spent around 7 hours on it and I feel good about it. In the first lesson, the teacher introduced most of the basic knowledge about AI and deep learning, their history, evolution, application, platforms, tools, materials, etc.
As a complete beginner, I have to say it’s not easy even for the first introductory course. There are a lot of new concepts, terms, apps, tools, and I have to go through them one by one.
For what I learned and remembered,
First, I know how deep learning is being applied in our real life, which includes Natural language processing (NLP), face recognition, speech recognition, and many others in this type that I am using in my life as well. Other applications including financial and logistical forecasting, text to speech are making our life much more convenient as well.
Before getting started with fast.ai, my understanding towards fastai was quite shallow and vague, now at least I know that it’s training processes for models. While giving input, different parameters should be set; after a result is shown, it needs to be compared with labels to have an idea of how well your model can give accurate results.
Also, what is so impressive for me is that I did run my first model and it was totally amazing. New terms I learned including: parallel distributed processing (PDP), GPU, weight assignment, weights, metric, loss, architecture, cell etc.
The software, including pytorch, fastai, and jupyter would still take me some time to learn and get familiar with.
Besides these, there is still a huge amount of new knowledge and information I need to learn and practice, fastai provides me with a great platform and a whole bunch of tools, materials, etc.