imaginAIry/imaginairy/modules/autoencoder.py

# pylama:ignore=W0613
import logging
from contextlib import contextmanager

import pytorch_lightning as pl
import torch
from torch.cuda import OutOfMemoryError

from imaginairy.feather_tile import rebuild_image, tile_image
from imaginairy.modules.diffusion.model import Decoder, Encoder
from imaginairy.modules.distributions import DiagonalGaussianDistribution
from imaginairy.modules.ema import LitEma
from imaginairy.utils import instantiate_from_config

logger = logging.getLogger(__name__)


class AutoencoderKL(pl.LightningModule):
    def __init__(
        self,
        ddconfig,
        lossconfig,
        embed_dim,
        ckpt_path=None,
        ignore_keys=None,
        image_key="image",
        colorize_nlabels=None,
        monitor=None,
        ema_decay=None,
        learn_logvar=False,
    ):
        super().__init__()
        self.learn_logvar = learn_logvar
        self.image_key = image_key
        self.encoder = Encoder(**ddconfig)
        self.decoder = Decoder(**ddconfig)
        self.loss = instantiate_from_config(lossconfig)
        assert ddconfig["double_z"]
        self.quant_conv = torch.nn.Conv2d(2 * ddconfig["z_channels"], 2 * embed_dim, 1)
        self.post_quant_conv = torch.nn.Conv2d(embed_dim, ddconfig["z_channels"], 1)
        self.embed_dim = embed_dim
        if colorize_nlabels is not None:
            assert isinstance(colorize_nlabels, int)
            self.register_buffer("colorize", torch.randn(3, colorize_nlabels, 1, 1))
        if monitor is not None:
            self.monitor = monitor

        self.use_ema = ema_decay is not None
        if self.use_ema:
            self.ema_decay = ema_decay
            assert 0.0 < ema_decay < 1.0
            self.model_ema = LitEma(self, decay=ema_decay)
            print(f"Keeping EMAs of {len(list(self.model_ema.buffers()))}.")

        if ckpt_path is not None:
            self.init_from_ckpt(ckpt_path, ignore_keys=ignore_keys)

    def init_from_ckpt(self, path, ignore_keys=None):
        sd = torch.load(path, map_location="cpu")["state_dict"]
        keys = list(sd.keys())
        ignore_keys = [] if ignore_keys is None else ignore_keys
        for k in keys:
            for ik in ignore_keys:
                if k.startswith(ik):
                    print(f"Deleting key {k} from state_dict.")
                    del sd[k]
        self.load_state_dict(sd, strict=False)
        print(f"Restored from {path}")

    @contextmanager
    def ema_scope(self, context=None):
        if self.use_ema:
            self.model_ema.store(self.parameters())
            self.model_ema.copy_to(self)
            if context is not None:
                print(f"{context}: Switched to EMA weights")
        try:
            yield None
        finally:
            if self.use_ema:
                self.model_ema.restore(self.parameters())
                if context is not None:
                    print(f"{context}: Restored training weights")

    def on_train_batch_end(self, *args, **kwargs):
        if self.use_ema:
            self.model_ema(self)

    def encode(self, x):
        h = self.encoder(x)
        moments = self.quant_conv(h)
        posterior = DiagonalGaussianDistribution(moments)
        return posterior

    def decode(self, z):
        try:
            return self.decode_all_at_once(z)
        except OutOfMemoryError:
            # Out of memory, trying sliced decoding.
            try:
                return self.decode_sliced(z, chunk_size=128)
            except OutOfMemoryError:
                return self.decode_sliced(z, chunk_size=64)

    def decode_all_at_once(self, z):
        z = self.post_quant_conv(z)
        dec = self.decoder(z)
        return dec

    def decode_sliced(self, z, chunk_size=128):
        """
        decodes the tensor in slices.

        This results in images that don't exactly match, so we overlap, feather, and merge to reduce
        (but not completely elminate) impact.
        """
        b, c, h, w = z.size()
        final_tensor = torch.zeros([1, 3, h * 8, w * 8], device=z.device)
        for z_latent in z.split(1):
            decoded_chunks = []
            overlap_pct = 0.5
            chunks = tile_image(
                z_latent, tile_size=chunk_size, overlap_percent=overlap_pct
            )

            for latent_chunk in chunks:
                latent_chunk = self.post_quant_conv(latent_chunk)
                dec = self.decoder(latent_chunk)
                decoded_chunks.append(dec)
            final_tensor = rebuild_image(
                decoded_chunks,
                base_img=final_tensor,
                tile_size=chunk_size * 8,
                overlap_percent=overlap_pct,
            )

            return final_tensor

    def forward(self, input, sample_posterior=True):  # noqa
        posterior = self.encode(input)
        if sample_posterior:
            z = posterior.sample()
        else:
            z = posterior.mode()
        dec = self.decode(z)
        return dec, posterior

    def get_input(self, batch, k):
        x = batch[k]
        if len(x.shape) == 3:
            x = x[..., None]
        x = x.permute(0, 3, 1, 2).to(memory_format=torch.contiguous_format).float()
        return x

    def training_step(self, batch, batch_idx, optimizer_idx):
        inputs = self.get_input(batch, self.image_key)
        reconstructions, posterior = self(inputs)

        if optimizer_idx == 0:
            # train encoder+decoder+logvar
            aeloss, log_dict_ae = self.loss(
                inputs,
                reconstructions,
                posterior,
                optimizer_idx,
                self.global_step,
                last_layer=self.get_last_layer(),
                split="train",
            )
            self.log(
                "aeloss",
                aeloss,
                prog_bar=True,
                logger=True,
                on_step=True,
                on_epoch=True,
            )
            self.log_dict(
                log_dict_ae, prog_bar=False, logger=True, on_step=True, on_epoch=False
            )
            return aeloss

        if optimizer_idx == 1:
            # train the discriminator
            discloss, log_dict_disc = self.loss(
                inputs,
                reconstructions,
                posterior,
                optimizer_idx,
                self.global_step,
                last_layer=self.get_last_layer(),
                split="train",
            )

            self.log(
                "discloss",
                discloss,
                prog_bar=True,
                logger=True,
                on_step=True,
                on_epoch=True,
            )
            self.log_dict(
                log_dict_disc, prog_bar=False, logger=True, on_step=True, on_epoch=False
            )
            return discloss
        return None

    def validation_step(self, batch, batch_idx):
        log_dict = self._validation_step(batch, batch_idx)
        with self.ema_scope():
            log_dict_ema = self._validation_step(  # noqa
                batch, batch_idx, postfix="_ema"
            )
        return log_dict

    def _validation_step(self, batch, batch_idx, postfix=""):
        inputs = self.get_input(batch, self.image_key)
        reconstructions, posterior = self(inputs)
        aeloss, log_dict_ae = self.loss(
            inputs,
            reconstructions,
            posterior,
            0,
            self.global_step,
            last_layer=self.get_last_layer(),
            split="val" + postfix,
        )

        discloss, log_dict_disc = self.loss(
            inputs,
            reconstructions,
            posterior,
            1,
            self.global_step,
            last_layer=self.get_last_layer(),
            split="val" + postfix,
        )

        self.log(f"val{postfix}/rec_loss", log_dict_ae[f"val{postfix}/rec_loss"])
        self.log_dict(log_dict_ae)
        self.log_dict(log_dict_disc)
        return self.log_dict

    def configure_optimizers(self):
        lr = self.learning_rate
        ae_params_list = (
            list(self.encoder.parameters())
            + list(self.decoder.parameters())
            + list(self.quant_conv.parameters())
            + list(self.post_quant_conv.parameters())
        )
        if self.learn_logvar:
            print(f"{self.__class__.__name__}: Learning logvar")
            ae_params_list.append(self.loss.logvar)
        opt_ae = torch.optim.Adam(ae_params_list, lr=lr, betas=(0.5, 0.9))
        opt_disc = torch.optim.Adam(
            self.loss.discriminator.parameters(), lr=lr, betas=(0.5, 0.9)
        )
        return [opt_ae, opt_disc], []

    def get_last_layer(self):
        return self.decoder.conv_out.weight

    @torch.no_grad()
    def log_images(self, batch, only_inputs=False, log_ema=False, **kwargs):
        log = {}
        x = self.get_input(batch, self.image_key)
        x = x.to(self.device)
        if not only_inputs:
            xrec, posterior = self(x)
            if x.shape[1] > 3:
                # colorize with random projection
                assert xrec.shape[1] > 3
                x = self.to_rgb(x)
                xrec = self.to_rgb(xrec)
            log["samples"] = self.decode(torch.randn_like(posterior.sample()))
            log["reconstructions"] = xrec
            if log_ema or self.use_ema:
                with self.ema_scope():
                    xrec_ema, posterior_ema = self(x)
                    if x.shape[1] > 3:
                        # colorize with random projection
                        assert xrec_ema.shape[1] > 3
                        xrec_ema = self.to_rgb(xrec_ema)
                    log["samples_ema"] = self.decode(
                        torch.randn_like(posterior_ema.sample())
                    )
                    log["reconstructions_ema"] = xrec_ema
        log["inputs"] = x
        return log

    # def to_rgb(self, x):
    #     assert self.image_key == "segmentation"
    #     if not hasattr(self, "colorize"):
    #         self.register_buffer("colorize", torch.randn(3, x.shape[1], 1, 1).to(x))
    #     x = F.conv2d(x, weight=self.colorize)
    #     x = 2.0 * (x - x.min()) / (x.max() - x.min()) - 1.0
    #     return x


class IdentityFirstStage(torch.nn.Module):
    def __init__(self, *args, vq_interface=False, **kwargs):
        self.vq_interface = vq_interface
        super().__init__()

    def encode(self, x, *args, **kwargs):
        return x

    def decode(self, x, *args, **kwargs):
        return x

    def quantize(self, x, *args, **kwargs):
        if self.vq_interface:
            return x, None, [None, None, None]
        return x

    def forward(self, x, *args, **kwargs):
        return x