app.ipynp

bear_type = 'grizzly', 'black', 'teddy'
path = Path('bears')

if not path.exists():
  # checks if the path already exists.
    path.mkdir()
    #When you call the mkdir() method on the path object, it creates a new directory with the name specified by path
    for o in bear_type:
      dest = (path/o)
        # makes a dest varibale that represents a path object for something in o, for example a "bear/grizzly" path
      dest.mkdir(exist_ok=True)
        # uses the mkdir method on the dest variable
      results = search_images_ddg(f'{o} bears')
        # goes and searches for urls with grizzly bears (or any thing in bear_type)
      download_images(dest, urls=results)
        # tells the computer where to save the images and what to download

fns = get_image_files(path)
fns

failed = verify_images(fns)
failed

failed.map(Path.unlink)

bears = DataBlock(
    # This specifies the types of the input and output data.
    blocks=(ImageBlock, CategoryBlock),
    # tells the datablock what files to get
    get_items=get_image_files,
    # selects 20% of the files for validation and seed insures that its the same 20%
    splitter=RandomSplitter(valid_pct=0.2, seed=42),
    # lparent_label is a function provided by fastai that simply gets the name of the folder a file is in. Because we put each of our bear images into folders based on the type of bear, this is going to give us the labels that we need.
    get_y=parent_label,
    # resizes each image
    item_tfms=Resize(128))

dls = bears.dataloaders(path)
dls.valid.show_batch(max_n=4, nrows=1)

bears = bears.new(
    item_tfms=RandomResizedCrop(224, min_scale=0.5),
    batch_tfms=aug_transforms())
dls = bears.dataloaders(path)

learn = vision_learner(dls, resnet18, metrics=error_rate)
learn.fine_tune(6)