cifar_mae_model.py

from functools import partial

import torch
import torch.nn as nn

from timm.models.vision_transformer import PatchEmbed, Block
from einops import repeat, rearrange


import numpy as np

import torch


def get_2d_sincos_pos_embed(embed_dim, grid_size, cls_token=False):
    """
    grid_size: int of the grid height and width
    return:
    pos_embed: [grid_size*grid_size, embed_dim] or [1+grid_size*grid_size, embed_dim] (w/ or w/o cls_token)
    """
    grid_h = np.arange(grid_size, dtype=np.float32)
    grid_w = np.arange(grid_size, dtype=np.float32)
    grid = np.meshgrid(grid_w, grid_h)  # here w goes first
    grid = np.stack(grid, axis=0)

    grid = grid.reshape([2, 1, grid_size, grid_size])
    pos_embed = get_2d_sincos_pos_embed_from_grid(embed_dim, grid)
    if cls_token:
        pos_embed = np.concatenate([np.zeros([1, embed_dim]), pos_embed], axis=0)
    return pos_embed


def get_2d_sincos_pos_embed_from_grid(embed_dim, grid):
    assert embed_dim % 2 == 0

    # use half of dimensions to encode grid_h
    emb_h = get_1d_sincos_pos_embed_from_grid(embed_dim // 2, grid[0])  # (H*W, D/2)
    emb_w = get_1d_sincos_pos_embed_from_grid(embed_dim // 2, grid[1])  # (H*W, D/2)

    emb = np.concatenate([emb_h, emb_w], axis=1) # (H*W, D)
    return emb


def get_1d_sincos_pos_embed_from_grid(embed_dim, pos):
    """
    embed_dim: output dimension for each position
    pos: a list of positions to be encoded: size (M,)
    out: (M, D)
    """
    assert embed_dim % 2 == 0
    omega = np.arange(embed_dim // 2, dtype=float)
    omega /= embed_dim / 2.
    omega = 1. / 10000**omega  # (D/2,)

    pos = pos.reshape(-1)  # (M,)
    out = np.einsum('m,d->md', pos, omega)  # (M, D/2), outer product

    emb_sin = np.sin(out) # (M, D/2)
    emb_cos = np.cos(out) # (M, D/2)

    emb = np.concatenate([emb_sin, emb_cos], axis=1)  # (M, D)
    return emb



def interpolate_pos_embed(model, checkpoint_model):
    
    if 'pos_embed' in checkpoint_model:
        print("yes")
        pos_embed_checkpoint = checkpoint_model['pos_embed']
        embedding_size = pos_embed_checkpoint.shape[-1]
        num_patches = model.patch_embed.num_patches
        num_extra_tokens = model.pos_embed.shape[-2] - num_patches
        # height (== width) for the checkpoint position embedding
        orig_size = int((pos_embed_checkpoint.shape[-2] - num_extra_tokens) ** 0.5)
        # height (== width) for the new position embedding
        new_size = int(num_patches ** 0.5)
        # class_token and dist_token are kept unchanged
        if orig_size != new_size:
            print("Position interpolate from %dx%d to %dx%d" % (orig_size, orig_size, new_size, new_size))
            extra_tokens = pos_embed_checkpoint[:, :num_extra_tokens]
            # only the position tokens are interpolated
            pos_tokens = pos_embed_checkpoint[:, num_extra_tokens:]
            pos_tokens = pos_tokens.reshape(-1, orig_size, orig_size, embedding_size).permute(0, 3, 1, 2)
            pos_tokens = torch.nn.functional.interpolate(
                pos_tokens, size=(new_size, new_size), mode='bicubic', align_corners=False)
            pos_tokens = pos_tokens.permute(0, 2, 3, 1).flatten(1, 2)
            new_pos_embed = torch.cat((extra_tokens, pos_tokens), dim=1)
            checkpoint_model['pos_embed'] = new_pos_embed




class HiddenLayer(nn.Module):
    def __init__(self, input_size, output_size):
        super(HiddenLayer, self).__init__()
        self.fc = nn.Linear(input_size, output_size)
        self.relu = nn.ReLU()

    def forward(self, x):
        x = self.fc(x)
        x = self.relu(x)
        return x

class MaskingModule(nn.Module):
    def __init__(self, in_features=16 * 16 * 768, out_features=16 * 16, hidden_size=512):
        """
        This is the masking module.
        :param in_features: Input dimension should match number of image patches
        :param out_features: Output dimension should match number of image patches dim(out)=dim(in)
        :param hidden_size: Size of hidden layer
        """
        super(MaskingModule, self).__init__()
        self.first_hidden_layer = HiddenLayer(in_features, hidden_size)
        self.output_layer = nn.Linear(hidden_size, out_features)

    def forward(self, x):
        x = torch.flatten(x, start_dim=1)
        x = self.first_hidden_layer(x)
        x = self.output_layer(x)
        return torch.sigmoid(x)


class MLOPatchMasking(nn.Module):
    # This replaces PatchShuffle with ranking probability and learnable masking module
    # returns unmasked patches, forward indices and backward indices.
    def __init__(self, in_features=16 * 16 * 768, out_features=16 * 16, hidden_size=512):
        super(MLOPatchMasking, self).__init__()
        self.first_hidden_layer = HiddenLayer(in_features, hidden_size)
        self.output_layer = nn.Linear(hidden_size, out_features)
    def forward(self, x, mask_ratio, random=False):
        N, L, D = x.shape  # batch, length, dim
        len_keep = int(L * (1 - mask_ratio))

        if random:
            mask_prob = torch.rand(N, L, device=x.device)  # noise in [0, 1]
            # print(mask_prob.shape)
            # exit()
        else:
            x2 = torch.flatten(x, start_dim=1)
            x2 = self.first_hidden_layer(x2)
            x2 = self.output_layer(x2)
            mask_prob = torch.sigmoid(x2)

        ids_shuffle = torch.argsort(mask_prob, dim=1)
        ids_restore = torch.argsort(ids_shuffle, dim=1)

        # print(x.shape)
        
        # keep the first subset
        ids_keep = ids_shuffle[:, :len_keep]


        # print(ids_keep.unsqueeze(-1).repeat(1, 1, D).shape)
        
        x_masked = torch.gather(x, dim=1, index=ids_keep.unsqueeze(-1).repeat(1, 1, D))

        # generate the binary mask: 0 is keep, 1 is remove
        mask = torch.ones([N, L], device=x.device)
        mask[:, :len_keep] = 0
        # unshuffle to get the binary mask
        mask = torch.gather(mask, dim=1, index=ids_restore)

        return x_masked, mask, ids_restore, mask_prob


class MLOPatchMasking2(nn.Module):
    # This replaces PatchShuffle with ranking probability and learnable masking module
    # returns unmasked patches, forward indices and backward indices.
    def __init__(self) -> None:
        super().__init__()
    def forward(self, x, mask_ratio, masking_module=None, random=False):
        N, L, D = x.shape  # batch, length, dim
        len_keep = int(L * (1 - mask_ratio))

        if random:
            mask_prob = torch.rand(N, L, device=x.device)  # noise in [0, 1]
            # print(mask_prob.shape)
            # exit()
        else:
            mask_prob = masking_module(x).to(x.device)
            # print(mask_prob.shape)
            # exit()
        ids_shuffle = torch.argsort(mask_prob, dim=1)
        ids_restore = torch.argsort(ids_shuffle, dim=1)

        # keep the first subset
        ids_keep = ids_shuffle[:, :len_keep]
        x_masked = torch.gather(x, dim=1, index=ids_keep.unsqueeze(-1).repeat(1, 1, D))

        # generate the binary mask: 0 is keep, 1 is remove
        mask = torch.ones([N, L], device=x.device)
        mask[:, :len_keep] = 0
        # unshuffle to get the binary mask
        mask = torch.gather(mask, dim=1, index=ids_restore)

        return x_masked, mask, ids_restore, mask_prob


class MaskedAutoencoderViT(nn.Module):
    """ Masked Autoencoder with VisionTransformer backbone
    """
    def __init__(self, img_size=32, patch_size=2, in_chans=3,
                 embed_dim=1024, depth=24, num_heads=16,
                 decoder_embed_dim=512, decoder_depth=8, decoder_num_heads=16,
                 mlp_ratio=4., norm_layer=nn.LayerNorm, norm_pix_loss=False):
        super().__init__()

        # --------------------------------------------------------------------------
        # MAE encoder specifics


        self.patch_embed = PatchEmbed(img_size, patch_size, in_chans, embed_dim)
        self.num_patches = self.patch_embed.num_patches
        self.embed_dim = embed_dim
        self.decoder_embed_dim = decoder_embed_dim

        self.cls_token = nn.Parameter(torch.zeros(1, 1, embed_dim))
        self.pos_embed = nn.Parameter(torch.zeros(1, self.num_patches + 1, embed_dim),
                                      requires_grad=True)  # fixed sin-cos embedding

        self.decoder_pos_embed = nn.Parameter(torch.zeros(1, self.num_patches + 1, decoder_embed_dim),
                                              requires_grad=True)  # fixed sin-cos embedding

        self.blocks = nn.ModuleList([
            Block(embed_dim, num_heads, mlp_ratio, qkv_bias=True, norm_layer=norm_layer)
            for i in range(depth)])
        self.norm = norm_layer(embed_dim)
        # --------------------------------------------------------------------------

        # --------------------------------------------------------------------------
        # MAE decoder specifics
        self.decoder_embed = nn.Linear(embed_dim, decoder_embed_dim, bias=True)
        # self.mask = MLOPatchMasking()

        self.mask_token = nn.Parameter(torch.zeros(1, 1, decoder_embed_dim))


        self.decoder_blocks = nn.ModuleList([
            Block(decoder_embed_dim, decoder_num_heads, mlp_ratio, qkv_bias=True, norm_layer=norm_layer)
            for i in range(decoder_depth)])

        self.decoder_norm = norm_layer(decoder_embed_dim)
        self.decoder_pred = nn.Linear(decoder_embed_dim, patch_size**2 * in_chans, bias=True) # decoder to patch
        # --------------------------------------------------------------------------

        self.norm_pix_loss = norm_pix_loss

        self.initialize_weights()

    def initialize_weights(self):
        # initialization
        # initialize (and freeze) pos_embed by sin-cos embedding
        pos_embed = get_2d_sincos_pos_embed(self.pos_embed.shape[-1], int(self.patch_embed.num_patches**.5), cls_token=True)
        self.pos_embed.data.copy_(torch.from_numpy(pos_embed).float().unsqueeze(0))

        decoder_pos_embed = get_2d_sincos_pos_embed(self.decoder_pos_embed.shape[-1], int(self.patch_embed.num_patches**.5), cls_token=True)
        self.decoder_pos_embed.data.copy_(torch.from_numpy(decoder_pos_embed).float().unsqueeze(0))

        # initialize patch_embed like nn.Linear (instead of nn.Conv2d)
        w = self.patch_embed.proj.weight.data
        torch.nn.init.xavier_uniform_(w.view([w.shape[0], -1]))

        # timm's trunc_normal_(std=.02) is effectively normal_(std=0.02) as cutoff is too big (2.)
        torch.nn.init.normal_(self.cls_token, std=.02)
        torch.nn.init.normal_(self.mask_token, std=.02)

        # initialize nn.Linear and nn.LayerNorm
        self.apply(self._init_weights)

    def _init_weights(self, m):
        if isinstance(m, nn.Linear):
            # we use xavier_uniform following official JAX ViT:
            torch.nn.init.xavier_uniform_(m.weight)
            if isinstance(m, nn.Linear) and m.bias is not None:
                nn.init.constant_(m.bias, 0)
        elif isinstance(m, nn.LayerNorm):
            nn.init.constant_(m.bias, 0)
            nn.init.constant_(m.weight, 1.0)

    def patchify(self, imgs):
        """
        imgs: (N, 3, H, W)
        x: (N, L, patch_size**2 *3)
        """
        p = self.patch_embed.patch_size[0]
        assert imgs.shape[2] == imgs.shape[3] and imgs.shape[2] % p == 0

        h = w = imgs.shape[2] // p
        x = imgs.reshape(shape=(imgs.shape[0], 3, h, p, w, p))
        x = torch.einsum('nchpwq->nhwpqc', x)
        x = x.reshape(shape=(imgs.shape[0], h * w, p**2 * 3))
        return x

    def unpatchify(self, x):
        """
        x: (N, L, patch_size**2 *3)
        imgs: (N, 3, H, W)
        """
        p = self.patch_embed.patch_size[0]
        h = w = int(x.shape[1]**.5)
        assert h * w == x.shape[1]
        
        x = x.reshape(shape=(x.shape[0], h, w, p, p, 3))
        x = torch.einsum('nhwpqc->nchpwq', x)
        imgs = x.reshape(shape=(x.shape[0], 3, h * p, h * p))
        return imgs

    def random_masking(self, x, mask_ratio):
        """
        Perform per-sample random masking by per-sample shuffling.
        Per-sample shuffling is done by argsort random noise.
        x: [N, L, D], sequence
        """
        N, L, D = x.shape  # batch, length, dim

        len_keep = int(L * (1 - mask_ratio))
        
        noise = torch.rand(N, L, device=x.device)  # noise in [0, 1]
        
        # sort noise for each sample
        ids_shuffle = torch.argsort(noise, dim=1)  # ascend: small is keep, large is remove
        ids_restore = torch.argsort(ids_shuffle, dim=1)

        # keep the first subset
        ids_keep = ids_shuffle[:, :len_keep]
        x_masked = torch.gather(x, dim=1, index=ids_keep.unsqueeze(-1).repeat(1, 1, D))

        # generate the binary mask: 0 is keep, 1 is remove
        mask = torch.ones([N, L], device=x.device)
        mask[:, :len_keep] = 0
        # unshuffle to get the binary mask
        mask = torch.gather(mask, dim=1, index=ids_restore)

        return x_masked, mask, ids_restore

    def forward_encoder_mlo(self, x, mask, ids_restore):
        # append cls token
        cls_token = self.cls_token + self.pos_embed[:, :1, :]
        cls_tokens = cls_token.expand(x.shape[0], -1, -1)
        x = torch.cat((cls_tokens, x), dim=1)

        # apply Transformer blocks
        for blk in self.blocks:
            x = blk(x)
        x2 = self.norm(x)

        return x2, mask, ids_restore

    def forward_encoder(self, x, mask_ratio=0):
        # embed patches
        # input of shape [B, 3, 224, 224]
        # patch of shape [B, T=(224/16)*(224/16)), 3 * 16 * 16] = [B, 196, 768]
        x = self.patch_embed(x)
        # add pos embed w/o cls token
        x = x + self.pos_embed[:, 1:, :]
        # masking: length -> length * mask_ratio
        # x of shape [B, (1-0.75)T,3 * 16 * 16] = [B, 49, 768]
        # mask of shape [B, T]
        # ids_restore of shape [B, T]
        x, mask, ids_restore = self.random_masking(x, mask_ratio)

        # append cls token
        cls_token = self.cls_token + self.pos_embed[:, :1, :]
        cls_tokens = cls_token.expand(x.shape[0], -1, -1)
        x = torch.cat((cls_tokens, x), dim=1)

        # apply Transformer blocks
        for blk in self.blocks:
            x = blk(x)
        x = self.norm(x)

        return x, mask, ids_restore

    def forward_decoder(self, x, ids_restore):
        # embed tokens
        x = self.decoder_embed(x)
        # append mask tokens to sequence ()
        # mask tokens is of shape mask_ratio * total patches
        mask_tokens = self.mask_token.repeat(x.shape[0], ids_restore.shape[1] + 1 - x.shape[1], 1)
        x_ = torch.cat([x[:, 1:, :], mask_tokens], dim=1)  # no cls token
        x_ = torch.gather(x_, dim=1, index=ids_restore.unsqueeze(-1).repeat(1, 1, x.shape[2]))  # unshuffle
        x = torch.cat([x[:, :1, :], x_], dim=1)  # append cls token

        # add pos embed
        x = x + self.decoder_pos_embed

        # apply Transformer blocks
        for blk in self.decoder_blocks:
            x = blk(x)
        x = self.decoder_norm(x)

        # predictor projection
        x = self.decoder_pred(x)

        # remove cls token
        x = x[:, 1:, :]

        return x

    def forward(self, imgs, mask, ids_restore):
        latent, mask, ids_restore = self.forward_encoder_mlo(imgs, mask, ids_restore)
        pred = self.forward_decoder(latent, ids_restore)  # [N, L, p*p*3]
        return pred


def mae_vit_base_patch16_dec512d8b(**kwargs):
    model = MaskedAutoencoderViT(
        patch_size=16, embed_dim=768, depth=12, num_heads=12,
        decoder_embed_dim=512, decoder_depth=8, decoder_num_heads=16,
        mlp_ratio=4, norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs)
    return model


def mae_vit_large_patch16_dec512d8b(**kwargs):
    model = MaskedAutoencoderViT(
        patch_size=16, embed_dim=1024, depth=24, num_heads=16,
        decoder_embed_dim=512, decoder_depth=8, decoder_num_heads=16,
        mlp_ratio=4, norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs)
    return model


def mae_vit_huge_patch14_dec512d8b(**kwargs):
    model = MaskedAutoencoderViT(
        patch_size=14, embed_dim=1280, depth=32, num_heads=16,
        decoder_embed_dim=512, decoder_depth=8, decoder_num_heads=16,
        mlp_ratio=4, norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs)
    return model





def cifar10_mae_vit_base_patch16_dec512d8b(**kwargs):
    model = MaskedAutoencoderViT(
        img_size=32, patch_size=2, embed_dim=768, depth=12, num_heads=12,
        decoder_embed_dim=512, decoder_depth=8, decoder_num_heads=16,
        mlp_ratio=4, norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs)
    return model

# set recommended archs
mae_vit_base_patch16 = mae_vit_base_patch16_dec512d8b  # decoder: 512 dim, 8 blocks
mae_vit_large_patch16 = mae_vit_large_patch16_dec512d8b  # decoder: 512 dim, 8 blocks
mae_vit_huge_patch14 = mae_vit_huge_patch14_dec512d8b  # decoder: 512 dim, 8 blocks

cifar10_mae_vit_base_patch16 = cifar10_mae_vit_base_patch16_dec512d8b  # decoder: 512 dim, 8 blocks