paddlepaddle (hereinafter referred to as fly paddle) is the author in the heart of the machine to see, when I just started road machine learning, have seen before Stanford Andrew Ng video machine learning, and b station is tensorflow learning video , as well as Mo Fan python and other learning videos, video brush several times, most will be able to change to change the data set to run other classification procedure, according to a dipper gourd painting feel difficult. Probably because the author non-computer professional, that no teacher's guide and a corresponding basis, want to achieve a lot of things do not know where to begin, learning materials do not know where to look. Fly paddle just to meet our group of "proper job" of students. Fly paddle contains a variety of information we need, the deepest impression of a course is "Baidu taught architect depth study", from the visual system inside and then to recommend to NLP, step by step, with our in-depth machine learning. And each stage will start from the most recent knowledge, before many do not understand something, or a smattering of knowledge, after listening to the teacher's class feel understood why. Data entry as a learning machine, fly paddle really is the best, I save a large amount of time searching for information. Programs do not understand, look for a direct flight paddle official website of the API, and then do not know to find the fly paddle forum, there are modules to solve the problem on GitHub, anyway, all kind of function complete. The above paddlehub includes pre-training model to meet the requirements of the usual learning. Fly paddle game is more like a small opportunity for students to participate in the project, after the game let me explain the benefit. Finally, thanks to what "Baidu taught architect depth study" of the complete course lecturer teacher, listening to his class so I really learned a soul machine learning, or the phrase "Everything lesson embedding".
