#hide
! [ -e /content ] && pip install -Uqq fastbook
import fastbook
fastbook.setup_book()

#hide
from fastbook import *
[[chapter_resnet]]
def get_data(url, presize, resize):
    path = untar_data(url)
    return DataBlock(
        blocks=(ImageBlock, CategoryBlock), get_items=get_image_files, 
        splitter=GrandparentSplitter(valid_name='val'),
        get_y=parent_label, item_tfms=Resize(presize),
        batch_tfms=[*aug_transforms(min_scale=0.5, size=resize),
                    Normalize.from_stats(*imagenet_stats)],
    ).dataloaders(path, bs=128)

dls = get_data(URLs.IMAGENETTE_160, 160, 128)

dls.show_batch(max_n=4)

def avg_pool(x): return x.mean((2,3))

def block(ni, nf): return ConvLayer(ni, nf, stride=2)
def get_model():
    return nn.Sequential(
        block(3, 16),
        block(16, 32),
        block(32, 64),
        block(64, 128),
        block(128, 256),
        nn.AdaptiveAvgPool2d(1),
        Flatten(),
        nn.Linear(256, dls.c))

def get_learner(m):
    return Learner(dls, m, loss_func=nn.CrossEntropyLoss(), metrics=accuracy
                  ).to_fp16()

learn = get_learner(get_model())

learn.lr_find()

learn.fit_one_cycle(5, 3e-3)

class ResBlock(Module):
    def __init__(self, ni, nf):
        self.convs = nn.Sequential(
            ConvLayer(ni,nf),
            ConvLayer(nf,nf, norm_type=NormType.BatchZero))
        
    def forward(self, x): return x + self.convs(x)

def _conv_block(ni,nf,stride):
    return nn.Sequential(
        ConvLayer(ni, nf, stride=stride),
        ConvLayer(nf, nf, act_cls=None, norm_type=NormType.BatchZero))

class ResBlock(Module):
    def __init__(self, ni, nf, stride=1):
        self.convs = _conv_block(ni,nf,stride)
        self.idconv = noop if ni==nf else ConvLayer(ni, nf, 1, act_cls=None)
        self.pool = noop if stride==1 else nn.AvgPool2d(2, ceil_mode=True)

    def forward(self, x):
        return F.relu(self.convs(x) + self.idconv(self.pool(x)))

def block(ni,nf): return ResBlock(ni, nf, stride=2)
learn = get_learner(get_model())

learn.fit_one_cycle(5, 3e-3)

def block(ni, nf):
    return nn.Sequential(ResBlock(ni, nf, stride=2), ResBlock(nf, nf))

learn = get_learner(get_model())
learn.fit_one_cycle(5, 3e-3)

def _resnet_stem(*sizes):
    return [
        ConvLayer(sizes[i], sizes[i+1], 3, stride = 2 if i==0 else 1)
            for i in range(len(sizes)-1)
    ] + [nn.MaxPool2d(kernel_size=3, stride=2, padding=1)]

#hide_output
_resnet_stem(3,32,32,64)

class ResNet(nn.Sequential):
    def __init__(self, n_out, layers, expansion=1):
        stem = _resnet_stem(3,32,32,64)
        self.block_szs = [64, 64, 128, 256, 512]
        for i in range(1,5): self.block_szs[i] *= expansion
        blocks = [self._make_layer(*o) for o in enumerate(layers)]
        super().__init__(*stem, *blocks,
                         nn.AdaptiveAvgPool2d(1), Flatten(),
                         nn.Linear(self.block_szs[-1], n_out))
    
    def _make_layer(self, idx, n_layers):
        stride = 1 if idx==0 else 2
        ch_in,ch_out = self.block_szs[idx:idx+2]
        return nn.Sequential(*[
            ResBlock(ch_in if i==0 else ch_out, ch_out, stride if i==0 else 1)
            for i in range(n_layers)
        ])

rn = ResNet(dls.c, [2,2,2,2])

learn = get_learner(rn)
learn.fit_one_cycle(5, 3e-3)

def _conv_block(ni,nf,stride):
    return nn.Sequential(
        ConvLayer(ni, nf//4, 1),
        ConvLayer(nf//4, nf//4, stride=stride), 
        ConvLayer(nf//4, nf, 1, act_cls=None, norm_type=NormType.BatchZero))

dls = get_data(URLs.IMAGENETTE_320, presize=320, resize=224)

rn = ResNet(dls.c, [3,4,6,3], 4)

learn = get_learner(rn)
learn.fit_one_cycle(20, 3e-3)