cwm/model/model_utils.py

import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F
from timm.models.layers import drop_path, to_2tuple


def _cfg(url='', **kwargs):
    return {
        'url': url,
        'num_classes': 400, 'input_size': (3, 224, 224), 'pool_size': None,
        'crop_pct': .9, 'interpolation': 'bicubic',
        'mean': (0.5, 0.5, 0.5), 'std': (0.5, 0.5, 0.5),
        **kwargs
    }


class DropPath(nn.Module):
    """Drop paths (Stochastic Depth) per sample  (when applied in main path of residual blocks).
    """
    def __init__(self, drop_prob=None):
        super(DropPath, self).__init__()
        self.drop_prob = drop_prob

    def forward(self, x):
        return drop_path(x, self.drop_prob, self.training)
    
    def extra_repr(self) -> str:
        return 'p={}'.format(self.drop_prob)


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)
        # commit this for the orignal BERT implement 
        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., attn_head_dim=None, flash_attention=False, k_bias=False, legacy=True, xla_flash=False):
        super().__init__()
        self.num_heads = num_heads
        head_dim = dim // num_heads
        if attn_head_dim is not None:
            head_dim = attn_head_dim
        all_head_dim = head_dim * self.num_heads
        self.scale = qk_scale or head_dim ** -0.5
        self.legacy = legacy

        self.xla_flash = xla_flash
        self.qkv = nn.Linear(dim, all_head_dim * 3, bias=False)

        if qkv_bias:
            self.q_bias = nn.Parameter(torch.zeros(all_head_dim))
            self.v_bias = nn.Parameter(torch.zeros(all_head_dim))
            if k_bias:
                self.k_bias = nn.Parameter(torch.zeros(all_head_dim))
            else:
                self.k_bias = None
        else:
            self.q_bias = None
            self.v_bias = None
            self.k_bias = None

        self.attn_drop = nn.Dropout(attn_drop)
        self.proj = nn.Linear(all_head_dim, dim)
        self.proj_drop = nn.Dropout(proj_drop)

    def forward(self, x):
        B, N, C = x.shape
        qkv_bias = None
        if self.q_bias is not None:
            if self.k_bias is not None:
                qkv_bias = torch.cat((self.q_bias, self.k_bias, self.v_bias))
            else:
                qkv_bias = torch.cat((self.q_bias, torch.zeros_like(self.v_bias, requires_grad=False), self.v_bias))

        qkv = F.linear(input=x, weight=self.qkv.weight, bias=qkv_bias)
        qkv = qkv.reshape(B, N, 3, self.num_heads, -1).permute(2, 0, 3, 1, 4)
        q, k, v = qkv[0], qkv[1], qkv[2]   # make torchscript happy (cannot use tensor as tuple)

        x = F.scaled_dot_product_attention(q, k, v, dropout_p=self.attn_drop.p)
        x = x.transpose(1, 2).reshape(B, N, -1)
        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., init_values=None, act_layer=nn.GELU, norm_layer=nn.LayerNorm,
                 attn_head_dim=None, in_dim=None, flash_attention=False, k_bias=False, legacy=False, xla_flash=False):
        super().__init__()
        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, attn_head_dim=attn_head_dim, flash_attention=flash_attention, k_bias=k_bias, legacy=legacy, xla_flash=xla_flash)
        # NOTE: drop path for stochastic depth, we shall see if this is better than dropout here
        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)

        if (init_values or 0) > 0:
            self.gamma_1 = nn.Parameter(init_values * torch.ones((dim)),requires_grad=True)
            self.gamma_2 = nn.Parameter(init_values * torch.ones((dim)),requires_grad=True)
        else:
            self.gamma_1, self.gamma_2 = None, None

    def forward(self, x):
        if self.gamma_1 is None:
            x = x + self.drop_path(self.attn(self.norm1(x)))
            x = x + self.drop_path(self.mlp(self.norm2(x)))
        else:
            x = x + self.drop_path(self.gamma_1 * self.attn(self.norm1(x)))
            x = x + self.drop_path(self.gamma_2 * self.mlp(self.norm2(x)))
        return x


class PatchEmbed(nn.Module):
    """ Image to Patch Embedding
    """
    def __init__(self, img_size=224, patch_size=(16, 16), in_chans=3, embed_dim=768, num_frames=16, tubelet_size=2):
        super().__init__()
        img_size = to_2tuple(img_size)

        self.tubelet_size = int(tubelet_size)
        if num_frames is not None:
            self.num_frames = int(num_frames)
            self.num_patches = (img_size[1] // patch_size[1]) * (img_size[0] // patch_size[0]) * (num_frames // self.tubelet_size)
        else:
            self.num_frames = None
            self.num_patches = None
        self.img_size = img_size
        self.patch_size = patch_size
        self.embed_dim = embed_dim

        self.proj = nn.Conv3d(in_channels=in_chans, out_channels=embed_dim, 
                            kernel_size = (self.tubelet_size,  patch_size[0],patch_size[1]), 
                            stride=(self.tubelet_size,  patch_size[0],  patch_size[1]))

    def forward(self, x, **kwargs):
        # B, C, T, H, W = x.shape
        # FIXME look at relaxing size constraints
        # assert H == self.img_size[0] and W == self.img_size[1], \
        #     f"Input image size ({H}*{W}) doesn't match model ({self.img_size[0]}*{self.img_size[1]})."
        x = self.proj(x).flatten(2).transpose(1, 2)
        return x
    
# sin-cos position encoding
# https://github.com/jadore801120/attention-is-all-you-need-pytorch/blob/master/transformer/Models.py#L31
def get_sinusoid_encoding_table(positions,
                                d_hid,
                                apply_sinusoid=True): 
    ''' Sinusoid position encoding table ''' 
    # TODO: make it with torch instead of numpy 
    def get_position_angle_vec(position): 
        return [position / np.power(10000, 2 * (hid_j // 2) / d_hid) for hid_j in range(d_hid)] 

    if isinstance(positions, int):
        sinusoid_table = np.array([get_position_angle_vec(pos_i) for pos_i in range(positions)])
    else:
        assert hasattr(positions, '__len__')
        sinusoid_table = np.array([get_position_angle_vec(pos_i) for pos_i in positions])
    if apply_sinusoid:
        sinusoid_table[:, 0::2] = np.sin(sinusoid_table[:, 0::2]) # dim 2i 
        sinusoid_table[:, 1::2] = np.cos(sinusoid_table[:, 1::2]) # dim 2i+1 

    return torch.FloatTensor(sinusoid_table).unsqueeze(0)