ibq_decoder.py

"""Minimal IBQ-Tokenizer decoder for Gestalt text-to-image generation.

Vendors only the decode path of the IBQ vision tokenizer
(TencentARC/IBQ-Tokenizer-16384, from the SEED-Voken codebase, Apache-2.0):
the VQGAN-style ``Decoder`` architecture plus the codebook embedding and
``post_quant_conv``. Weights are loaded from the published Lightning
checkpoint, preferring the EMA parameters (the authors decode under
``ema_scope()``).

Decode semantics match SEED-Voken exactly:
  indices (b, h*w)
    -> get_codebook_entry(indices, shape=(b, h, w, c))
    -> post_quant_conv -> decoder -> clamp(-1, 1) -> PIL image.
"""

import numpy as np
import torch
import torch.nn as nn


def nonlinearity(x):
    # swish
    return x * torch.sigmoid(x)


def Normalize(in_channels):
    return torch.nn.GroupNorm(num_groups=32, num_channels=in_channels, eps=1e-6, affine=True)


class Upsample(nn.Module):
    def __init__(self, in_channels, with_conv):
        super().__init__()
        self.with_conv = with_conv
        if self.with_conv:
            self.conv = torch.nn.Conv2d(in_channels, in_channels, kernel_size=3, stride=1, padding=1)

    def forward(self, x):
        x = torch.nn.functional.interpolate(x, scale_factor=2.0, mode="nearest")
        if self.with_conv:
            x = self.conv(x)
        return x


class ResnetBlock(nn.Module):
    def __init__(self, *, in_channels, out_channels=None, conv_shortcut=False,
                 dropout, temb_channels=512):
        super().__init__()
        self.in_channels = in_channels
        out_channels = in_channels if out_channels is None else out_channels
        self.out_channels = out_channels
        self.use_conv_shortcut = conv_shortcut

        self.norm1 = Normalize(in_channels)
        self.conv1 = torch.nn.Conv2d(in_channels, out_channels, kernel_size=3, stride=1, padding=1)
        if temb_channels > 0:
            self.temb_proj = torch.nn.Linear(temb_channels, out_channels)
        self.norm2 = Normalize(out_channels)
        self.dropout = torch.nn.Dropout(dropout)
        self.conv2 = torch.nn.Conv2d(out_channels, out_channels, kernel_size=3, stride=1, padding=1)
        if self.in_channels != self.out_channels:
            if self.use_conv_shortcut:
                self.conv_shortcut = torch.nn.Conv2d(in_channels, out_channels, kernel_size=3, stride=1, padding=1)
            else:
                self.nin_shortcut = torch.nn.Conv2d(in_channels, out_channels, kernel_size=1, stride=1, padding=0)

    def forward(self, x, temb):
        h = x
        h = self.norm1(h)
        h = nonlinearity(h)
        h = self.conv1(h)

        if temb is not None:
            h = h + self.temb_proj(nonlinearity(temb))[:, :, None, None]

        h = self.norm2(h)
        h = nonlinearity(h)
        h = self.dropout(h)
        h = self.conv2(h)

        if self.in_channels != self.out_channels:
            if self.use_conv_shortcut:
                x = self.conv_shortcut(x)
            else:
                x = self.nin_shortcut(x)

        return x + h


class AttnBlock(nn.Module):
    def __init__(self, in_channels):
        super().__init__()
        self.in_channels = in_channels

        self.norm = Normalize(in_channels)
        self.q = torch.nn.Conv2d(in_channels, in_channels, kernel_size=1, stride=1, padding=0)
        self.k = torch.nn.Conv2d(in_channels, in_channels, kernel_size=1, stride=1, padding=0)
        self.v = torch.nn.Conv2d(in_channels, in_channels, kernel_size=1, stride=1, padding=0)
        self.proj_out = torch.nn.Conv2d(in_channels, in_channels, kernel_size=1, stride=1, padding=0)

    def forward(self, x):
        h_ = x
        h_ = self.norm(h_)
        q = self.q(h_)
        k = self.k(h_)
        v = self.v(h_)

        # compute attention
        b, c, h, w = q.shape
        q = q.reshape(b, c, h * w)
        q = q.permute(0, 2, 1)   # b, hw, c
        k = k.reshape(b, c, h * w)  # b, c, hw
        w_ = torch.bmm(q, k)     # b, hw, hw    w[b,i,j]=sum_c q[b,i,c]k[b,c,j]
        w_ = w_ * (int(c) ** (-0.5))
        w_ = torch.nn.functional.softmax(w_, dim=2)

        # attend to values
        v = v.reshape(b, c, h * w)
        w_ = w_.permute(0, 2, 1)   # b, hw, hw (first hw of k, second of q)
        h_ = torch.bmm(v, w_)     # b, c, hw (hw of q) h_[b,c,j] = sum_i v[b,c,i] w_[b,i,j]
        h_ = h_.reshape(b, c, h, w)

        h_ = self.proj_out(h_)

        return x + h_


class Decoder(nn.Module):
    def __init__(self, *, ch, out_ch, ch_mult=(1, 2, 4, 8), num_res_blocks,
                 attn_resolutions, dropout=0.0, resamp_with_conv=True, in_channels,
                 resolution, z_channels, give_pre_end=False, **ignorekwargs):
        super().__init__()
        self.ch = ch
        self.temb_ch = 0
        self.num_resolutions = len(ch_mult)
        self.num_res_blocks = num_res_blocks
        self.resolution = resolution
        self.in_channels = in_channels
        self.give_pre_end = give_pre_end

        # compute in_ch_mult, block_in and curr_res at lowest res
        in_ch_mult = (1,) + tuple(ch_mult)
        block_in = ch * ch_mult[self.num_resolutions - 1]
        curr_res = resolution // 2 ** (self.num_resolutions - 1)
        self.z_shape = (1, z_channels, curr_res, curr_res)

        # z to block_in
        self.conv_in = torch.nn.Conv2d(z_channels, block_in, kernel_size=3, stride=1, padding=1)

        # middle
        self.mid = nn.Module()
        self.mid.block_1 = ResnetBlock(in_channels=block_in, out_channels=block_in,
                                       temb_channels=self.temb_ch, dropout=dropout)
        self.mid.attn_1 = AttnBlock(block_in)
        self.mid.block_2 = ResnetBlock(in_channels=block_in, out_channels=block_in,
                                       temb_channels=self.temb_ch, dropout=dropout)

        # upsampling
        self.up = nn.ModuleList()
        for i_level in reversed(range(self.num_resolutions)):
            block = nn.ModuleList()
            attn = nn.ModuleList()
            block_out = ch * ch_mult[i_level]
            for i_block in range(self.num_res_blocks + 1):
                block.append(ResnetBlock(in_channels=block_in, out_channels=block_out,
                                         temb_channels=self.temb_ch, dropout=dropout))
                block_in = block_out
                if curr_res in attn_resolutions:
                    attn.append(AttnBlock(block_in))
            up = nn.Module()
            up.block = block
            up.attn = attn
            if i_level != 0:
                up.upsample = Upsample(block_in, resamp_with_conv)
                curr_res = curr_res * 2
            self.up.insert(0, up)  # prepend to get consistent order

        # end
        self.norm_out = Normalize(block_in)
        self.conv_out = torch.nn.Conv2d(block_in, out_ch, kernel_size=3, stride=1, padding=1)

    def forward(self, z):
        # timestep embedding (unused, unconditional decoder)
        temb = None

        # z to block_in
        h = self.conv_in(z)

        # middle
        h = self.mid.block_1(h, temb)
        h = self.mid.attn_1(h)
        h = self.mid.block_2(h, temb)

        # upsampling
        for i_level in reversed(range(self.num_resolutions)):
            for i_block in range(self.num_res_blocks + 1):
                h = self.up[i_level].block[i_block](h, temb)
                if len(self.up[i_level].attn) > 0:
                    h = self.up[i_level].attn[i_block](h)
            if i_level != 0:
                h = self.up[i_level].upsample(h)

        # end
        h = self.norm_out(h)
        h = nonlinearity(h)
        h = self.conv_out(h)
        return h


class IBQDecoder(nn.Module):
    """Codebook embedding + post_quant_conv + VQGAN decoder for IBQ tokens.

    Matches ``TencentARC/IBQ-Tokenizer-16384`` (imagenet256_16384.ckpt,
    n_embed=16384, embed_dim=256, z_channels=256, 256x256 native resolution).
    """

    DDCONFIG = dict(
        double_z=False,
        z_channels=256,
        resolution=256,
        in_channels=3,
        out_ch=3,
        ch=128,
        ch_mult=[1, 1, 2, 2, 4],
        num_res_blocks=4,
        attn_resolutions=[16],
        dropout=0.0,
    )

    def __init__(self):
        super().__init__()
        cfg = self.DDCONFIG
        self.decoder = Decoder(**cfg)
        self.quantize_embedding = nn.Embedding(16384, 256)
        self.post_quant_conv = nn.Conv2d(256, cfg["z_channels"], 1)

    def load_from_checkpoint(self, ckpt_path: str):
        """Load EMA weights (fallback: plain) from the Lightning .ckpt."""
        sd = torch.load(ckpt_path, map_location="cpu", weights_only=False)["state_dict"]
        target_keys = set(self.state_dict().keys())
        new_params = {}
        ema_key_map = {}
        # First pass: build EMA-name map ("model_ema.<dots-removed>" -> plain name)
        for k in sd.keys():
            s_name = k.replace(".", "")
            if k.startswith("model_ema."):
                ema_key_map[s_name] = k
        # Second pass: prefer EMA values, fall back to plain values
        for k in target_keys:
            if k.startswith("quantize_embedding."):
                base = "quantize.embedding." + k.split(".", 1)[1]
            else:
                base = k
            ema_name = "model_ema." + base.replace(".", "")
            if ema_name in sd:
                new_params[k] = sd[ema_name]
            elif base in sd:
                new_params[k] = sd[base]
        missing = target_keys - set(new_params.keys())
        if missing:
            raise ValueError(f"IBQ checkpoint missing keys: {sorted(missing)[:5]}")
        self.load_state_dict(new_params, strict=True)
        return self

    def decode_indices(self, indices: torch.Tensor, lat_h: int = 32, lat_w: int = 32) -> torch.Tensor:
        """IBQ codebook indices (b, h*w) -> decoded image tensor (b, 3, H, W) in [-1, 1].

        Mirrors ``IndexPropagationQuantize.get_codebook_entry`` +
        ``VQModel.decode`` from SEED-Voken.
        """
        b = indices.shape[0]
        c = self.quantize_embedding.embedding_dim
        z_q = self.quantize_embedding(indices)          # (b, h*w, c)
        z_q = z_q.view(b, lat_h, lat_w, c)              # shape=(b, h, w, c)
        z_q = z_q.permute(0, 3, 1, 2).contiguous()      # (b, c, h, w)
        quant = self.post_quant_conv(z_q)
        return self.decoder(quant)

    @staticmethod
    def to_pil(x: torch.Tensor) -> "Image.Image":
        """SEED-Voken ``custom_to_pil``: [-1,1] CHW tensor -> PIL RGB image."""
        from PIL import Image
        x = x.detach().cpu()
        x = torch.clamp(x, -1.0, 1.0)
        x = (x + 1.0) / 2.0
        x = x.permute(1, 2, 0).numpy()
        x = (255 * x).astype(np.uint8)
        return Image.fromarray(x)