Lesson1.ipynb(lecture -1) doubts


(Nandamuri Hari Naga Sumanth) #1

I am unable to understand/(I want to understand better) some stuff from lesson1.ipynb notebook please help me in this.

  1. Why is sz parameter assigned value 224 ?
  2. I know tfms means transforms but what is exactly is happening in this line (tfms=tfms_from_model(arch, sz)), why is parameter arch passed ?
    3)learn = ConvLearner.pretrained(arch, data, precompute=True), since precompute=True are we using precomputed weights trained on Imagenet data?, If yes then what is learn.fit() method doing ?.
    These questions may be basic but please help me.

(Amrit ) #2

@HariSumanth9 please view my notes at the link below. They may shed some light to some of your questions. There are also plenty of other great forum threads for additional information sources. Happy Learning.