models/timesformer.py

# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
# Copyright 2020 Ross Wightman
# Modified Model definition

import torch
import torch.nn as nn
from functools import partial
import math
import warnings
import torch.nn.functional as F
import numpy as np



from timesformer.models.vit_utils import IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD
from timesformer.models.helpers import load_pretrained
from timesformer.models.vit_utils import DropPath, to_2tuple, trunc_normal_

# from timesformer.models.build import MODEL_REGISTRY
from torch import einsum
from einops import rearrange, reduce, repeat
import torchvision

def _cfg(url='', **kwargs):
    return {
        'url': url,
        'num_classes': 1000, 'input_size': (3, 224, 224), 'pool_size': None,
        'crop_pct': .9, 'interpolation': 'bicubic',
        'mean': IMAGENET_DEFAULT_MEAN, 'std': IMAGENET_DEFAULT_STD,
        'first_conv': 'patch_embed.proj', 'classifier': 'head',
        **kwargs
    }


default_cfgs = {
    'timesformer_vit_base_patch16_224': _cfg(
        url='https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-vitjx/jx_vit_base_p16_224-80ecf9dd.pth',
        mean=(0.5, 0.5, 0.5), std=(0.5, 0.5, 0.5),
    ),
}

class Mlp(nn.Module):
    def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
        super().__init__()
        out_features = out_features or in_features
        hidden_features = hidden_features or in_features
        self.fc1 = nn.Linear(in_features, hidden_features)
        self.act = act_layer()
        self.fc2 = nn.Linear(hidden_features, out_features)
        self.drop = nn.Dropout(drop)

    def forward(self, x):
        x = self.fc1(x)
        x = self.act(x)
        x = self.drop(x)
        x = self.fc2(x)
        x = self.drop(x)
        return x

class Attention(nn.Module):
    def __init__(self, dim, num_heads=8, qkv_bias=False, qk_scale=None, attn_drop=0., proj_drop=0., with_qkv=True):
        super().__init__()
        self.num_heads = num_heads
        head_dim = dim // num_heads
        self.scale = qk_scale or head_dim ** -0.5
        self.with_qkv = with_qkv
        if self.with_qkv:
           self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)
           self.proj = nn.Linear(dim, dim)
           self.proj_drop = nn.Dropout(proj_drop)
        self.attn_drop = nn.Dropout(attn_drop)

    def forward(self, x):
        B, N, C = x.shape
        if self.with_qkv:
           qkv = self.qkv(x).reshape(B, N, 3, self.num_heads, C // self.num_heads).permute(2, 0, 3, 1, 4)
           q, k, v = qkv[0], qkv[1], qkv[2]
        else:
           qkv = x.reshape(B, N, self.num_heads, C // self.num_heads).permute(0, 2, 1, 3)
           q, k, v  = qkv, qkv, qkv

        attn = (q @ k.transpose(-2, -1)) * self.scale
        attn = attn.softmax(dim=-1)
        attn = self.attn_drop(attn)

        x = (attn @ v).transpose(1, 2).reshape(B, N, C)
        if self.with_qkv:
           x = self.proj(x)
           x = self.proj_drop(x)
        return x

class Block(nn.Module):

    def __init__(self, dim, num_heads, mlp_ratio=4., qkv_bias=False, qk_scale=None, drop=0., attn_drop=0.,
                 drop_path=0.1, act_layer=nn.GELU, norm_layer=nn.LayerNorm, attention_type='divided_space_time'):
        super().__init__()
        self.attention_type = attention_type
        assert(attention_type in ['divided_space_time', 'space_only','joint_space_time'])

        self.norm1 = norm_layer(dim)
        self.attn = Attention(
           dim, num_heads=num_heads, qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop)

        ## Temporal Attention Parameters
        if self.attention_type == 'divided_space_time':
            self.temporal_norm1 = norm_layer(dim)
            self.temporal_attn = Attention(
              dim, num_heads=num_heads, qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop)
            self.temporal_fc = nn.Linear(dim, dim)

        ## drop path
        self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
        self.norm2 = norm_layer(dim)
        mlp_hidden_dim = int(dim * mlp_ratio)
        self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)


    def forward(self, x, B, T, W):
        num_spatial_tokens = (x.size(1) - 1) // T
        H = num_spatial_tokens // W

        if self.attention_type in ['space_only', 'joint_space_time']:
            x = x + self.drop_path(self.attn(self.norm1(x)))
            x = x + self.drop_path(self.mlp(self.norm2(x)))
            return x
        elif self.attention_type == 'divided_space_time':
            ## Temporal
            xt = x[:,1:,:]
            xt = rearrange(xt, 'b (h w t) m -> (b h w) t m',b=B,h=H,w=W,t=T)
            res_temporal = self.drop_path(self.temporal_attn(self.temporal_norm1(xt)))
            res_temporal = rearrange(res_temporal, '(b h w) t m -> b (h w t) m',b=B,h=H,w=W,t=T)
            res_temporal = self.temporal_fc(res_temporal)
            xt = x[:,1:,:] + res_temporal

            ## Spatial
            init_cls_token = x[:,0,:].unsqueeze(1)
            cls_token = init_cls_token.repeat(1, T, 1)
            cls_token = rearrange(cls_token, 'b t m -> (b t) m',b=B,t=T).unsqueeze(1)
            xs = xt
            xs = rearrange(xs, 'b (h w t) m -> (b t) (h w) m',b=B,h=H,w=W,t=T)
            xs = torch.cat((cls_token, xs), 1)
            res_spatial = self.drop_path(self.attn(self.norm1(xs)))

            ### Taking care of CLS token
            cls_token = res_spatial[:,0,:]
            cls_token = rearrange(cls_token, '(b t) m -> b t m',b=B,t=T)
            cls_token = torch.mean(cls_token,1,True) ## averaging for every frame
            res_spatial = res_spatial[:,1:,:]
            res_spatial = rearrange(res_spatial, '(b t) (h w) m -> b (h w t) m',b=B,h=H,w=W,t=T)
            res = res_spatial
            x = xt

            ## Mlp
            x = torch.cat((init_cls_token, x), 1) + torch.cat((cls_token, res), 1)
            x = x + self.drop_path(self.mlp(self.norm2(x)))
            return x

class PatchEmbed(nn.Module):
    """ Image to Patch Embedding
    """
    def __init__(self, img_size=224, patch_size=16, in_chans=3, embed_dim=768):
        super().__init__()
        img_size = to_2tuple(img_size)
        patch_size = to_2tuple(patch_size)
        num_patches = (img_size[1] // patch_size[1]) * (img_size[0] // patch_size[0])
        self.img_size = img_size
        self.patch_size = patch_size
        self.num_patches = num_patches

        self.proj = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=patch_size)

    def forward(self, x):
        B, C, T, H, W = x.shape
        x = rearrange(x, 'b c t h w -> (b t) c h w')
        x = self.proj(x)
        W = x.size(-1)
        x = x.flatten(2).transpose(1, 2)
        return x, T, W


class VisionTransformer(nn.Module):
    """ Vision Transformere
    """
    def __init__(self, img_size=224, patch_size=16, in_chans=3, num_classes=1000, embed_dim=768, depth=12,
                 num_heads=12, mlp_ratio=4., qkv_bias=False, qk_scale=None, drop_rate=0., attn_drop_rate=0.,
                 drop_path_rate=0.1, hybrid_backbone=None, norm_layer=nn.LayerNorm, 
                 num_frames=8, attention_type='divided_space_time', dropout=0., 
                 return_hidden_state=False):
        super().__init__()
        self.attention_type = attention_type
        self.depth = depth
        self.dropout = nn.Dropout(dropout)
        self.num_classes = num_classes
        self.num_features = self.embed_dim = embed_dim  # num_features for consistency with other models
        self.patch_embed = PatchEmbed(
            img_size=img_size, patch_size=patch_size, in_chans=in_chans, embed_dim=embed_dim)
        num_patches = self.patch_embed.num_patches

        ## Positional Embeddings
        self.cls_token = nn.Parameter(torch.zeros(1, 1, embed_dim))
        self.pos_embed = nn.Parameter(torch.zeros(1, num_patches+1, embed_dim))
        self.pos_drop = nn.Dropout(p=drop_rate)
        if self.attention_type != 'space_only':
            self.time_embed = nn.Parameter(torch.zeros(1, num_frames, embed_dim))
            self.time_drop = nn.Dropout(p=drop_rate)

        ## Attention Blocks
        dpr = [x.item() for x in torch.linspace(0, drop_path_rate, self.depth)]  # stochastic depth decay rule
        self.blocks = nn.ModuleList([
            Block(
                dim=embed_dim, num_heads=num_heads, mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
                drop=drop_rate, attn_drop=attn_drop_rate, drop_path=dpr[i], norm_layer=norm_layer, attention_type=self.attention_type)
            for i in range(self.depth)])
        self.norm = norm_layer(embed_dim)

        # Classifier head
        self.head = nn.Linear(embed_dim, num_classes) if num_classes > 0 else nn.Identity()

        trunc_normal_(self.pos_embed, std=.02)
        trunc_normal_(self.cls_token, std=.02)
        self.apply(self._init_weights)

        ## initialization of temporal attention weights
        if self.attention_type == 'divided_space_time':
            i = 0
            for m in self.blocks.modules():
                m_str = str(m)
                if 'Block' in m_str:
                    if i > 0:
                      nn.init.constant_(m.temporal_fc.weight, 0)
                      nn.init.constant_(m.temporal_fc.bias, 0)
                    i += 1
        print("Load custom timesformer")

        self.return_hidden_state = return_hidden_state

    def _init_weights(self, m):
        if isinstance(m, nn.Linear):
            trunc_normal_(m.weight, std=.02)
            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)

    @torch.jit.ignore
    def no_weight_decay(self):
        return {'pos_embed', 'cls_token', 'time_embed'}

    def get_classifier(self):
        return self.head

    def reset_classifier(self, num_classes, global_pool=''):
        self.num_classes = num_classes
        self.head = nn.Linear(self.embed_dim, num_classes) if num_classes > 0 else nn.Identity()

    def forward_features(self, x, attention_type=None):
        all_hidden_states = () if self.return_hidden_state else None
        B = x.shape[0]
        x, T, W = self.patch_embed(x)
        cls_tokens = self.cls_token.expand(x.size(0), -1, -1)
        x = torch.cat((cls_tokens, x), dim=1)

        ## resizing the positional embeddings in case they don't match the input at inference
        if x.size(1) != self.pos_embed.size(1):
            pos_embed = self.pos_embed
            cls_pos_embed = pos_embed[0,0,:].unsqueeze(0).unsqueeze(1)
            other_pos_embed = pos_embed[0,1:,:].unsqueeze(0).transpose(1, 2)
            P = int(other_pos_embed.size(2) ** 0.5)
            H = x.size(1) // W
            other_pos_embed = other_pos_embed.reshape(1, x.size(2), P, P)
            new_pos_embed = F.interpolate(other_pos_embed, size=(H, W), mode='nearest')
            new_pos_embed = new_pos_embed.flatten(2)
            new_pos_embed = new_pos_embed.transpose(1, 2)
            new_pos_embed = torch.cat((cls_pos_embed, new_pos_embed), 1)
            x = x + new_pos_embed
        else:
            x = x + self.pos_embed
        x = self.pos_drop(x)

        if attention_type is None:
            attention_type = self.attention_type
        ## Time Embeddings
        if attention_type != 'space_only':
            cls_tokens = x[:B, 0, :].unsqueeze(1)
            x = x[:,1:]
            x = rearrange(x, '(b t) n m -> (b n) t m',b=B,t=T)
            ## Resizing time embeddings in case they don't match
            if T != self.time_embed.size(1):
                time_embed = self.time_embed.transpose(1, 2)
                new_time_embed = F.interpolate(time_embed, size=(T), mode='nearest')
                new_time_embed = new_time_embed.transpose(1, 2)
                x = x + new_time_embed
            else:
                x = x + self.time_embed
            x = self.time_drop(x)
            x = rearrange(x, '(b n) t m -> b (n t) m',b=B,t=T)
            x = torch.cat((cls_tokens, x), dim=1)

        ## Attention blocks
        for blk in self.blocks:
            x = blk(x, B, T, W)

            if self.return_hidden_state:
                all_hidden_states = all_hidden_states + (self.norm(x),)

        ### Predictions for space-only baseline
        if attention_type == 'space_only':
            x = rearrange(x, '(b t) n m -> b t n m',b=B,t=T)
            x = torch.mean(x, 1) # averaging predictions for every frame

        x = self.norm(x)

        if self.return_hidden_state:
            return x, all_hidden_states
        else:
            return x

    def forward(self, x, attention_type=None):
        x = self.forward_features(x, attention_type=attention_type)
        return x

def _conv_filter(state_dict, patch_size=16):
    """ convert patch embedding weight from manual patchify + linear proj to conv"""
    out_dict = {}
    for k, v in state_dict.items():
        if 'patch_embed.proj.weight' in k:
            if v.shape[-1] != patch_size:
                patch_size = v.shape[-1]
            v = v.reshape((v.shape[0], 3, patch_size, patch_size))
        out_dict[k] = v
    return out_dict

# @MODEL_REGISTRY.register()
class timesformer_vit_base_patch16_224(nn.Module):
    def __init__(self, cfg, **kwargs):
        super(timesformer_vit_base_patch16_224, self).__init__()
        self.pretrained=True
        patch_size = 16
        self.model = VisionTransformer(img_size=cfg.DATA.TRAIN_CROP_SIZE, 
            num_classes=cfg.MODEL.NUM_CLASSES, patch_size=patch_size, 
            embed_dim=768, depth=12, num_heads=12, mlp_ratio=4, qkv_bias=True, norm_layer=partial(nn.LayerNorm, eps=1e-6), drop_rate=0., attn_drop_rate=0., drop_path_rate=0.1, num_frames=cfg.DATA.NUM_FRAMES, attention_type=cfg.TIMESFORMER.ATTENTION_TYPE, **kwargs)

        self.attention_type = cfg.TIMESFORMER.ATTENTION_TYPE
        self.model.default_cfg = default_cfgs['timesformer_vit_base_patch16_224']
        self.num_patches = (cfg.DATA.TRAIN_CROP_SIZE // patch_size) * (cfg.DATA.TRAIN_CROP_SIZE // patch_size)
        pretrained_model=cfg.TIMESFORMER.PRETRAINED_MODEL
        if self.pretrained:
            load_pretrained(self.model, num_classes=self.model.num_classes, in_chans=kwargs.get('in_chans', 3), filter_fn=_conv_filter, img_size=cfg.DATA.TRAIN_CROP_SIZE, num_patches=self.num_patches, attention_type=self.attention_type, pretrained_model=pretrained_model)

    def forward(self, x):
        x = self.model(x)
        return x

# @MODEL_REGISTRY.register()
class TimeSformer(nn.Module):
    def __init__(self, img_size=224, patch_size=16, num_classes=400, num_frames=8, 
        attention_type='divided_space_time',  embed_dim=768, pretrained_model='', 
        audio_as_image=True, space_only_for_images=False,  **kwargs):
        super(TimeSformer, self).__init__()
        self.pretrained=True
        self.model = VisionTransformer(img_size=img_size, num_classes=num_classes, 
            patch_size=patch_size, embed_dim=embed_dim, depth=12, num_heads=12, mlp_ratio=4, 
            qkv_bias=True, norm_layer=partial(nn.LayerNorm, eps=1e-6), 
            drop_rate=0., attn_drop_rate=0., drop_path_rate=0.1, num_frames=num_frames, 
            attention_type=attention_type,  **kwargs)
        self.num_frames = num_frames
        self.embed_dim = embed_dim
        self.attention_type = attention_type
        self.model.default_cfg = default_cfgs['timesformer_vit_base_patch'+str(patch_size)+'_224']
        self.num_patches = (img_size // patch_size) * (img_size // patch_size)
        self.img_size = img_size
        self.audio_as_image = audio_as_image
        self.space_only_for_images = space_only_for_images
        if self.pretrained:
            load_pretrained(self.model, num_classes=self.model.num_classes, in_chans=kwargs.get('in_chans', 3), filter_fn=_conv_filter, img_size=img_size, num_frames=num_frames, num_patches=self.num_patches, attention_type=self.attention_type, pretrained_model=pretrained_model)
    def forward(self, x, external_features=None):
        
        if x.ndim == 4: # image as input
            if not self.space_only_for_images:
                x = x.unsqueeze(2).expand(-1, -1, self.num_frames, -1, -1) # B, C, f, H, W
            else:
                x = x.unsqueeze(2) # B, C, f, H, W
        if x.ndim == 3: # image as input
            B, H, W = x.shape
            if self.audio_as_image:
                x = x.unsqueeze(1).expand(-1, 3, -1, -1) # add C channel 
                x = torchvision.transforms.functional.resize(x, (self.img_size, self.img_size))
                if not self.space_only_for_images:
                    x = x.unsqueeze(2).expand(-1, -1, self.num_frames, -1, -1) # B, C, f, H, W
                else:
                    x = x.unsqueeze(2) # B, C, f, H, W
            else:
                if H != W: # audio (1024, 128) another option is to make a square image
                    if H > W:
                        w = H/self.num_frames
                        # if w > W: 
                        #     w = w/2
                        a = w - W # overlap a < 0
                        x = x.unfold(1, W, int(W+a)) # from mel to square frames 
                        x = x.unsqueeze(1).expand(-1, 3, -1, -1, -1) # add C channel 
                        x = [torchvision.transforms.functional.resize(x[:, :, i, :, :], (self.img_size, self.img_size)).unsqueeze(2) for i in range(self.num_frames)]
                        x = torch.cat(x, dim=2)

            x = (x - x.min()) / (x.max() - x.min()) # does not help
        # print(x.shape)
        if x.shape[-3] == 1 and self.space_only_for_images: # space only for images 
            attention_type = 'space_only'
        else:
            attention_type = None
        x = self.model(x, attention_type=attention_type)
        
        
        return x