src/models/focus_resnet_module.py

from typing import Any, List

import torch
import torch.nn.functional as F
from torch import nn
from pytorch_lightning import LightningModule
from torchmetrics import MaxMetric, MeanAbsoluteError, MinMetric
from torchmetrics.classification.accuracy import Accuracy
import torchvision.models as models


class ResNetLitModule(LightningModule):
    def __init__(
        self,
        resnet_type: str = "resnet18",
        pretrained=False,
        lr: float = 0.001,
        weight_decay: float = 0.0005,
    ):
        """Initialize function for a resnet module.

        Args:
            resnet_type (str, optional): Type of the used resnet network. Defaults to
                "ResNet".
                Can be one of the following values: "resnet18",
                    "resnet34", "resnet50", "resnet101", "resnet152", "resnext50_32x4d",
                    "resnext101_32x8d", "wide_resnet50_2", "wide_resnet101_2"
            pretrained (bool, optional): if True loads pytorch pretrained models.
                Defaults to False.
        """
        super().__init__()

        # this line allows to access init params with 'self.hparams' attribute
        # it also ensures init params will be stored in ckpt
        self.save_hyperparameters(logger=False)

        # loss function
        self.criterion = torch.nn.MSELoss()

        # use separate metric instance for train, val and test step
        # to ensure a proper reduction over the epoch
        self.train_mae = MeanAbsoluteError()
        self.val_mae = MeanAbsoluteError()
        self.test_mae = MeanAbsoluteError()

        # for logging best so far validation accuracy
        self.val_mae_best = MinMetric()

        self.pretrained = pretrained

        if resnet_type == "resnet18":
            resnet_constructor = models.resnet18
        elif resnet_type == "resnet34":
            resnet_constructor = models.resnet34
        elif resnet_type == "resnet50":
            resnet_constructor = models.resnet50
        elif resnet_type == "resnet101":
            resnet_constructor = models.resnet101
        elif resnet_type == "resnet152":
            resnet_constructor = models.resnet152
        elif resnet_type == "resnext50_32x4d":
            resnet_constructor = models.resnext50_32x4d
        elif resnet_type == "resnext101_32x8d":
            resnet_constructor = models.resnext101_32x8d
        elif resnet_type == "wide_resnet50_2":
            resnet_constructor = models.wide_resnet50_2
        elif resnet_type == "wide_resnet101_2":
            resnet_constructor = models.wide_resnet101_2
        else:
            raise Exception(f"did not find model type: {resnet_type}")

        backbone = resnet_constructor(pretrained=pretrained)
        # init a pretrained resnet

        num_filters = backbone.fc.in_features
        layers = list(backbone.children())[:-1]
        self.feature_extractor = nn.Sequential(*layers)

        self.fc = nn.Linear(num_filters, 1)

    def forward(self, x):
        representations = self.feature_extractor(x).flatten(1)
        x = self.fc(representations)
        return x

    def step(self, batch: Any):
        x = batch["image"]
        y = batch["focus_height"]
        logits = self.forward(x)
        loss = self.criterion(logits, y.unsqueeze(1))
        preds = torch.squeeze(logits)
        return loss, preds, y

    def training_step(self, batch: Any, batch_idx: int):
        loss, preds, targets = self.step(batch)

        # log train metrics
        mae = self.train_mae(preds, targets)
        self.log("train/loss", loss, on_step=False, on_epoch=True, prog_bar=False)
        self.log("train/mae", mae, on_step=False, on_epoch=True, prog_bar=True)

        # we can return here dict with any tensors
        # and then read it in some callback or in `training_epoch_end()`` below
        # remember to always return loss from `training_step()` or else
        # backpropagation will fail!
        return {"loss": loss, "preds": preds, "targets": targets}

    def training_epoch_end(self, outputs: List[Any]):
        # `outputs` is a list of dicts returned from `training_step()`
        pass

    def validation_step(self, batch: Any, batch_idx: int):
        loss, preds, targets = self.step(batch)

        # log val metrics
        mae = self.val_mae(preds, targets)
        self.log("val/loss", loss, on_step=False, on_epoch=True, prog_bar=False)
        self.log("val/mae", mae, on_step=False, on_epoch=True, prog_bar=True)

        return {"loss": loss, "preds": preds, "targets": targets}

    def validation_epoch_end(self, outputs: List[Any]):
        mae = self.val_mae.compute()  # get val accuracy from current epoch
        self.val_mae_best.update(mae)
        self.log(
            "val/mae_best", self.val_mae_best.compute(), on_epoch=True, prog_bar=True
        )

    def test_step(self, batch: Any, batch_idx: int):
        loss, preds, targets = self.step(batch)

        # log test metrics
        mae = self.test_mae(preds, targets)
        self.log("test/loss", loss, on_step=False, on_epoch=True)
        self.log("test/mae", mae, on_step=False, on_epoch=True)

    def test_epoch_end(self, outputs: List[Any]):
        print(outputs)
        pass

    def on_epoch_end(self):
        # reset metrics at the end of every epoch
        self.train_mae.reset()
        self.test_mae.reset()
        self.val_mae.reset()

    def configure_optimizers(self):
        """Choose what optimizers and learning-rate schedulers.

        Normally you'd need one. But in the case of GANs or similar you might
        have multiple.

        See examples here:
            https://pytorch-lightning.readthedocs.io/en/latest/common/lightning_module.html#configure-optimizers
        """
        return torch.optim.Adam(
            params=self.parameters(),
            lr=self.hparams.lr,
            weight_decay=self.hparams.weight_decay,
        )