From 5c4dea983b1fad3d2ca209063fc21e51bd17d845 Mon Sep 17 00:00:00 2001 From: triple-Mu Date: Tue, 5 Jul 2022 18:25:52 +0800 Subject: [PATCH] add convit --- flowvision/models/__init__.py | 4 +- flowvision/models/convit.py | 546 ++++++++++++++++++++++++++++++++++ 2 files changed, 547 insertions(+), 3 deletions(-) create mode 100644 flowvision/models/convit.py diff --git a/flowvision/models/__init__.py b/flowvision/models/__init__.py index ac493283..264a5ce8 100644 --- a/flowvision/models/__init__.py +++ b/flowvision/models/__init__.py @@ -32,9 +32,7 @@ from .van import * from .levit import * from .mobilevit import * -from .deit_III import * -from .cait import * -from .dla import * +from .convit import * from . import style_transfer from . import detection diff --git a/flowvision/models/convit.py b/flowvision/models/convit.py new file mode 100644 index 00000000..91c3c903 --- /dev/null +++ b/flowvision/models/convit.py @@ -0,0 +1,546 @@ +import oneflow as flow +import oneflow.nn as nn +from functools import partial +import oneflow.nn.functional as F + + +from flowvision.models.helpers import to_2tuple +from flowvision.layers import trunc_normal_, DropPath, Mlp +from .registry import ModelCreator +from .utils import load_state_dict_from_url + +__all__ = [ + "VisionTransformer", + "convit_tiny", + "convit_small", + "convit_base", +] + +model_urls = { + "convit_tiny": "https://oneflow-public.oss-cn-beijing.aliyuncs.com/model_zoo/flowvision/classification/ConVit/convit_tiny.zip", + "convit_small": "https://oneflow-public.oss-cn-beijing.aliyuncs.com/model_zoo/flowvision/classification/ConVit/convit_small.zip", + "convit_base": "https://oneflow-public.oss-cn-beijing.aliyuncs.com/model_zoo/flowvision/classification/ConVit/convit_base.zip", +} + + +class GPSA(nn.Module): + def __init__( + self, + dim, + num_heads=8, + qkv_bias=False, + qk_scale=None, + attn_drop=0.0, + proj_drop=0.0, + locality_strength=1.0, + use_local_init=True, + ): + super().__init__() + self.num_heads = num_heads + self.dim = dim + head_dim = dim // num_heads + self.scale = qk_scale or head_dim ** -0.5 + + self.qk = nn.Linear(dim, dim * 2, bias=qkv_bias) + self.v = nn.Linear(dim, dim, bias=qkv_bias) + + self.attn_drop = nn.Dropout(attn_drop) + self.proj = nn.Linear(dim, dim) + self.pos_proj = nn.Linear(3, num_heads) + self.proj_drop = nn.Dropout(proj_drop) + self.locality_strength = locality_strength + self.gating_param = nn.Parameter(flow.ones(self.num_heads)) + self.apply(self._init_weights) + if use_local_init: + self.local_init(locality_strength=locality_strength) + + def _init_weights(self, m): + if isinstance(m, nn.Linear): + trunc_normal_(m.weight, std=0.02) + 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) + + def forward(self, x): + B, N, C = x.shape + if not hasattr(self, "rel_indices") or self.rel_indices.size(1) != N: + self.get_rel_indices(N) + + attn = self.get_attention(x) + v = ( + self.v(x) + .reshape(B, N, self.num_heads, C // self.num_heads) + .permute(0, 2, 1, 3) + ) + x = (attn @ v).transpose(1, 2).reshape(B, N, C) + x = self.proj(x) + x = self.proj_drop(x) + return x + + def get_attention(self, x): + B, N, C = x.shape + qk = ( + self.qk(x) + .reshape(B, N, 2, self.num_heads, C // self.num_heads) + .permute(2, 0, 3, 1, 4) + ) + q, k = qk[0], qk[1] + pos_score = self.rel_indices.expand(B, -1, -1, -1) + pos_score = self.pos_proj(pos_score).permute(0, 3, 1, 2) + patch_score = (q @ k.transpose(-2, -1)) * self.scale + patch_score = patch_score.softmax(dim=-1) + pos_score = pos_score.softmax(dim=-1) + + gating = self.gating_param.view(1, -1, 1, 1) + attn = (1.0 - flow.sigmoid(gating)) * patch_score + flow.sigmoid( + gating + ) * pos_score + attn /= attn.sum(dim=-1).unsqueeze(-1) + attn = self.attn_drop(attn) + return attn + + def get_attention_map(self, x, return_map=False): + + attn_map = self.get_attention(x).mean(0) # average over batch + distances = self.rel_indices.squeeze()[:, :, -1] ** 0.5 + dist = flow.einsum("nm,hnm->h", (distances, attn_map)) + dist /= distances.size(0) + if return_map: + return dist, attn_map + else: + return dist + + def local_init(self, locality_strength=1.0): + + self.v.weight.data.copy_(flow.eye(self.dim)) + locality_distance = 1 # max(1,1/locality_strength**.5) + + kernel_size = int(self.num_heads ** 0.5) + center = (kernel_size - 1) / 2 if kernel_size % 2 == 0 else kernel_size // 2 + for h1 in range(kernel_size): + for h2 in range(kernel_size): + position = h1 + kernel_size * h2 + self.pos_proj.weight.data[position, 2] = -1 + self.pos_proj.weight.data[position, 1] = ( + 2 * (h1 - center) * locality_distance + ) + self.pos_proj.weight.data[position, 0] = ( + 2 * (h2 - center) * locality_distance + ) + self.pos_proj.weight.data *= locality_strength + + def get_rel_indices(self, num_patches): + img_size = int(num_patches ** 0.5) + rel_indices = flow.zeros(1, num_patches, num_patches, 3) + ind = flow.arange(img_size, dtype=flow.float32).view(1, -1) - flow.arange( + img_size, dtype=flow.float32 + ).view(-1, 1) + indx = ind.repeat(img_size, img_size) + indy = ind.repeat_interleave(img_size, dim=0).repeat_interleave(img_size, dim=1) + indd = indx ** 2 + indy ** 2 + rel_indices[:, :, :, 2] = indd.unsqueeze(0) + rel_indices[:, :, :, 1] = indy.unsqueeze(0) + rel_indices[:, :, :, 0] = indx.unsqueeze(0) + device = self.qk.weight.device + self.rel_indices = rel_indices.to(device) + + +class MHSA(nn.Module): + def __init__( + self, + dim, + num_heads=8, + qkv_bias=False, + qk_scale=None, + attn_drop=0.0, + proj_drop=0.0, + ): + super().__init__() + self.num_heads = num_heads + head_dim = dim // num_heads + self.scale = qk_scale or head_dim ** -0.5 + + self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias) + self.attn_drop = nn.Dropout(attn_drop) + self.proj = nn.Linear(dim, dim) + self.proj_drop = nn.Dropout(proj_drop) + self.apply(self._init_weights) + + def _init_weights(self, m): + if isinstance(m, nn.Linear): + trunc_normal_(m.weight, std=0.02) + 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) + + def get_attention_map(self, x, return_map=False): + B, N, C = x.shape + 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] + attn_map = (q @ k.transpose(-2, -1)) * self.scale + attn_map = attn_map.softmax(dim=-1).mean(0) + + img_size = int(N ** 0.5) + ind = flow.arange(img_size).view(1, -1) - flow.arange(img_size).view(-1, 1) + indx = ind.repeat(img_size, img_size) + indy = ind.repeat_interleave(img_size, dim=0).repeat_interleave(img_size, dim=1) + indd = indx ** 2 + indy ** 2 + distances = indd ** 0.5 + distances = distances.to("cuda") + + dist = flow.einsum("nm,hnm->h", (distances, attn_map)) + dist /= N + + if return_map: + return dist, attn_map + else: + return dist + + def forward(self, x): + B, N, C = x.shape + 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] + + 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) + x = self.proj(x) + x = self.proj_drop(x) + return x + + +class Block(nn.Module): + def __init__( + self, + dim, + num_heads, + mlp_ratio=4.0, + qkv_bias=False, + qk_scale=None, + drop=0.0, + attn_drop=0.0, + drop_path=0.0, + act_layer=nn.GELU, + norm_layer=nn.LayerNorm, + use_gpsa=True, + **kwargs, + ): + super().__init__() + self.norm1 = norm_layer(dim) + self.use_gpsa = use_gpsa + if self.use_gpsa: + self.attn = GPSA( + dim, + num_heads=num_heads, + qkv_bias=qkv_bias, + qk_scale=qk_scale, + attn_drop=attn_drop, + proj_drop=drop, + **kwargs, + ) + else: + self.attn = MHSA( + dim, + num_heads=num_heads, + qkv_bias=qkv_bias, + qk_scale=qk_scale, + attn_drop=attn_drop, + proj_drop=drop, + **kwargs, + ) + self.drop_path = DropPath(drop_path) if drop_path > 0.0 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): + x = x + self.drop_path(self.attn(self.norm1(x))) + x = x + self.drop_path(self.mlp(self.norm2(x))) + return x + + +class PatchEmbed(nn.Module): + """ Image to Patch Embedding, from timm + """ + + 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 + ) + self.apply(self._init_weights) + + def forward(self, x): + B, C, H, W = x.shape + 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 + + def _init_weights(self, m): + if isinstance(m, nn.Linear): + trunc_normal_(m.weight, std=0.02) + 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) + + +class HybridEmbed(nn.Module): + """ CNN Feature Map Embedding, from timm + """ + + def __init__( + self, backbone, img_size=224, feature_size=None, in_chans=3, embed_dim=768 + ): + super().__init__() + assert isinstance(backbone, nn.Module) + img_size = to_2tuple(img_size) + self.img_size = img_size + self.backbone = backbone + if feature_size is None: + with flow.no_grad(): + training = backbone.training + if training: + backbone.eval() + o = self.backbone(flow.zeros(1, in_chans, img_size[0], img_size[1]))[-1] + feature_size = o.shape[-2:] + feature_dim = o.shape[1] + backbone.train(training) + else: + feature_size = to_2tuple(feature_size) + feature_dim = self.backbone.feature_info.channels()[-1] + self.num_patches = feature_size[0] * feature_size[1] + self.proj = nn.Linear(feature_dim, embed_dim) + self.apply(self._init_weights) + + def forward(self, x): + x = self.backbone(x)[-1] + x = x.flatten(2).transpose(1, 2) + x = self.proj(x) + return x + + +class VisionTransformer(nn.Module): + """ Vision Transformer with support for patch or hybrid CNN input stage + """ + + def __init__( + self, + img_size=224, + patch_size=16, + in_chans=3, + num_classes=1000, + embed_dim=48, + depth=12, + num_heads=12, + mlp_ratio=4.0, + qkv_bias=False, + qk_scale=None, + drop_rate=0.0, + attn_drop_rate=0.0, + drop_path_rate=0.0, + hybrid_backbone=None, + norm_layer=nn.LayerNorm, + global_pool=None, + local_up_to_layer=10, + locality_strength=1.0, + use_pos_embed=True, + ): + super().__init__() + self.num_classes = num_classes + self.local_up_to_layer = local_up_to_layer + self.num_features = ( + self.embed_dim + ) = embed_dim # num_features for consistency with other models + self.locality_strength = locality_strength + self.use_pos_embed = use_pos_embed + + if hybrid_backbone is not None: + self.patch_embed = HybridEmbed( + hybrid_backbone, + img_size=img_size, + in_chans=in_chans, + embed_dim=embed_dim, + ) + else: + 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 + self.num_patches = num_patches + + self.cls_token = nn.Parameter(flow.zeros(1, 1, embed_dim)) + self.pos_drop = nn.Dropout(p=drop_rate) + + if self.use_pos_embed: + self.pos_embed = nn.Parameter(flow.zeros(1, num_patches, embed_dim)) + trunc_normal_(self.pos_embed, std=0.02) + + dpr = [ + x.item() for x in flow.linspace(0, drop_path_rate, 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, + use_gpsa=True, + locality_strength=locality_strength, + ) + if i < local_up_to_layer + else 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, + use_gpsa=False, + ) + for i in range(depth) + ] + ) + self.norm = norm_layer(embed_dim) + + # Classifier head + self.feature_info = [dict(num_chs=embed_dim, reduction=0, module="head")] + self.head = ( + nn.Linear(embed_dim, num_classes) if num_classes > 0 else nn.Identity() + ) + + trunc_normal_(self.cls_token, std=0.02) + self.head.apply(self._init_weights) + + def _init_weights(self, m): + if isinstance(m, nn.Linear): + trunc_normal_(m.weight, std=0.02) + 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) + + def no_weight_decay(self): + return {"pos_embed", "cls_token"} + + 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): + B = x.shape[0] + x = self.patch_embed(x) + + cls_tokens = self.cls_token.expand(B, -1, -1) + + if self.use_pos_embed: + x = x + self.pos_embed + x = self.pos_drop(x) + + for u, blk in enumerate(self.blocks): + if u == self.local_up_to_layer: + x = flow.cat((cls_tokens, x), dim=1) + x = blk(x) + + x = self.norm(x) + return x[:, 0] + + def forward(self, x): + x = self.forward_features(x) + x = self.head(x) + return x + + +@ModelCreator.register_model +def convit_tiny(pretrained: bool = False, progress: bool = True, **kwargs): + num_heads = 4 + kwargs.setdefault("embed_dim", 48) + kwargs["embed_dim"] *= num_heads + model = VisionTransformer( + num_heads=num_heads, norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs + ) + if pretrained: + state_dict = load_state_dict_from_url( + model_urls["convit_tiny"], model_dir="./checkpoints", progress=progress, + ) + model.load_state_dict(state_dict) + return model + + +@ModelCreator.register_model +def convit_small(pretrained: bool = False, progress: bool = True, **kwargs): + num_heads = 9 + kwargs.setdefault("embed_dim", 48) + kwargs["embed_dim"] *= num_heads + model = VisionTransformer( + num_heads=num_heads, norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs + ) + if pretrained: + state_dict = load_state_dict_from_url( + model_urls["convit_small"], model_dir="./checkpoints", progress=progress, + ) + model.load_state_dict(state_dict) + return model + + +@ModelCreator.register_model +def convit_base(pretrained: bool = False, progress: bool = True, **kwargs): + num_heads = 16 + kwargs.setdefault("embed_dim", 48) + kwargs["embed_dim"] *= num_heads + model = VisionTransformer( + num_heads=num_heads, norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs + ) + if pretrained: + state_dict = load_state_dict_from_url( + model_urls["convit_base"], model_dir="./checkpoints", progress=progress, + ) + model.load_state_dict(state_dict) + return model