features/edgevit3 (#214)

* add backbone model 'EdgeVit'

configs/classification/imagenet/edgevit/EdgeVit_b512x8_300e_jpg.py

@@ -0,0 +1,111 @@
+_base_ = '../../../base.py'
+
+log_config = dict(
+    interval=10,
+    hooks=[dict(type='TextLoggerHook'),
+           dict(type='TensorboardLoggerHook')])
+
+# model settings
+model = dict(
+    type='Classification',
+    backbone=dict(
+        type='EdgeVit',
+        depth=[1, 1, 3, 2],
+        embed_dim=[36, 72, 144, 288],
+        head_dim=36,
+        mlp_ratio=[4] * 4,
+        qkv_bias=True,
+        num_classes=1000,
+        drop_path_rate=0.1,
+        sr_ratios=[4, 2, 2, 1]),
+    head=dict(
+        type='ClsHead',
+        with_avg_pool=True,
+        in_channels=288,
+        loss_config=dict(
+            type='CrossEntropyLossWithLabelSmooth', label_smooth=0),
+    ))
+
+data_train_list = 'data/imagenet_raw/meta/train_labeled.txt'
+data_train_root = 'data/imagenet_raw/train/'
+data_test_list = 'data/imagenet_raw/meta/val_labeled.txt'
+data_test_root = 'data/imagenet_raw/validation/'
+
+dataset_type = 'ClsDataset'
+img_norm_cfg = dict(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+train_pipeline = [
+    dict(
+        type='MAEFtAugment',
+        input_size=224,
+        color_jitter=0.4,
+        auto_augment='rand-m9-mstd0.5-inc1',
+        interpolation='bicubic',
+        re_prob=0.25,
+        re_mode='pixel',
+        re_count=1,
+        mean=(0.485, 0.456, 0.406),
+        std=(0.229, 0.224, 0.225),
+        is_train=True),
+]
+test_pipeline = [
+    dict(
+        type='MAEFtAugment',
+        input_size=224,
+        mean=(0.485, 0.456, 0.406),
+        std=(0.229, 0.224, 0.225),
+        is_train=False,
+    ),
+]
+
+data = dict(
+    imgs_per_gpu=512,
+    workers_per_gpu=10,
+    use_repeated_augment_sampler=True,
+    train=dict(
+        type=dataset_type,
+        data_source=dict(
+            list_file=data_train_list,
+            root=data_train_root,
+            type='ClsSourceImageList'),
+        pipeline=train_pipeline),
+    val=dict(
+        type=dataset_type,
+        data_source=dict(
+            list_file=data_test_list,
+            root=data_test_root,
+            type='ClsSourceImageList'),
+        pipeline=test_pipeline))
+
+eval_config = dict(initial=True, interval=1, gpu_collect=True)
+eval_pipelines = [
+    dict(
+        mode='test',
+        data=data['val'],
+        dist_eval=True,
+        evaluators=[dict(type='ClsEvaluator', topk=(1, 5))],
+    )
+]
+
+# additional hooks
+custom_hooks = []
+
+# optimizer
+optimizer = dict(type='AdamW', lr=2e-3, weight_decay=0.05)
+
+# learning policy
+lr_config = dict(
+    policy='CosineAnnealingWarmupByEpoch',
+    min_lr=1e-5,
+    warmup='linear',
+    warmup_iters=5,
+    warmup_ratio=1e-6,
+    # warmup_lr=1e-6,
+    warmup_by_epoch=True,
+    by_epoch=True)
+
+checkpoint_config = dict(interval=10)
+
+# runtime settings
+total_epochs = 300
+
+ema = dict(decay=0.99996)

configs/classification/imagenet/edgevit/imagenet_edgeVIT_s_jpg.py

@@ -0,0 +1,42 @@
+_base_ = './EdgeVit_b512x8_300e_jpg.py'
+# model settings
+model = dict(
+    type='Classification',
+    train_preprocess=['mixUp'],
+    mixup_cfg=dict(
+        mixup_alpha=0.8,
+        cutmix_alpha=1.0,
+        cutmix_minmax=None,
+        prob=1.0,
+        switch_prob=0.5,
+        mode='batch',
+        label_smoothing=0.1,
+        num_classes=1000),
+    backbone=dict(
+        type='EdgeVit',
+        depth=[1, 2, 5, 3],
+        embed_dim=[48, 96, 240, 384],
+        head_dim=48,
+        mlp_ratio=[4] * 4,
+        qkv_bias=True,
+        num_classes=1000,
+        drop_path_rate=0.1,
+        sr_ratios=[4, 2, 2, 1]),
+    head=dict(
+        type='ClsHead',
+        with_avg_pool=True,
+        in_channels=384,
+        loss_config={
+            'type': 'SoftTargetCrossEntropy',
+        },
+        with_fc=True))
+
+data = dict(
+    imgs_per_gpu=128,
+    workers_per_gpu=10,
+    use_repeated_augment_sampler=True,
+)
+
+# optimizer
+update_interval = 8
+optimizer_config = dict(update_interval=update_interval)

configs/classification/imagenet/edgevit/imagenet_edgeVIT_xs_jpg.py

@@ -0,0 +1,31 @@
+_base_ = './EdgeVit_b512x8_300e_jpg.py'
+model = dict(
+    type='Classification',
+    backbone=dict(
+        type='EdgeVit',
+        depth=[1, 1, 3, 1],
+        embed_dim=[48, 96, 240, 384],
+        head_dim=48,
+        mlp_ratio=[4] * 4,
+        qkv_bias=True,
+        num_classes=1000,
+        drop_path_rate=0.1,
+        sr_ratios=[4, 2, 2, 1]),
+    head=dict(
+        type='ClsHead',
+        with_avg_pool=True,
+        in_channels=384,
+        loss_config=dict(
+            type='CrossEntropyLossWithLabelSmooth', label_smooth=0.1),
+    ))
+
+# input data settings
+data = dict(
+    imgs_per_gpu=256,
+    workers_per_gpu=10,
+    use_repeated_augment_sampler=True,
+)
+
+# optimizer
+update_interval = 4
+optimizer_config = dict(update_interval=update_interval)

configs/classification/imagenet/edgevit/imagenet_edgeVIT_xxs_jpg.py

@@ -0,0 +1,32 @@
+_base_ = './EdgeVit_b512x8_300e_jpg.py'
+# model settings
+model = dict(
+    type='Classification',
+    backbone=dict(
+        type='EdgeVit',
+        depth=[1, 1, 3, 2],
+        embed_dim=[36, 72, 144, 288],
+        head_dim=36,
+        mlp_ratio=[4] * 4,
+        qkv_bias=True,
+        num_classes=1000,
+        drop_path_rate=0.1,
+        sr_ratios=[4, 2, 2, 1]),
+    head=dict(
+        type='ClsHead',
+        with_avg_pool=True,
+        in_channels=288,
+        loss_config=dict(
+            type='CrossEntropyLossWithLabelSmooth', label_smooth=0.1),
+    ))
+
+# input data settings
+data = dict(
+    imgs_per_gpu=512,
+    workers_per_gpu=10,
+    use_repeated_augment_sampler=True,
+)
+
+# optimizer
+update_interval = 2
+optimizer_config = dict(update_interval=update_interval)

docs/source/model_zoo_cls.md

@@ -82,5 +82,8 @@
 | efficientformer_l1 | [efficientformer_l1](https://github.com/alibaba/EasyCV/tree/master/configs/classification/imagenet/efficientformer/efficientformer_l1.py) | 80.102   | 94.934   | 1820    | 7.5    | [model](https://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/EasyCV/modelzoo/classification/efficientformer/efficientformer_l1_1000d.pth) |
 | efficientformer_l3 | [efficientformer_l3](https://github.com/alibaba/EasyCV/tree/master/configs/classification/imagenet/efficientformer/efficientformer_l3.py) | 82.272   | 96.028   | 2436    | 13.07    | [model](https://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/EasyCV/modelzoo/classification/efficientformer/efficientformer_l3_300d.pth) |
 | efficientformer_l7 | [efficientformer_l7](https://github.com/alibaba/EasyCV/tree/master/configs/classification/imagenet/efficientformer/efficientformer_l7.py) | 83.076   | 96.44   | 1622    | 18.96    | [model](https://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/EasyCV/modelzoo/classification/efficientformer/efficientformer_l7_300d.pth) |
+| EdgeVit_xxs_b512_224 | [EdgeVit_xxs_b512_224](https://github.com/alibaba/EasyCV/tree/master/configs/classification/imagenet/edgevit/imagenet_edgeVIT_xxs_jpg.py) | 75.18   | 92.188   | 206   | 8.67    | [model](http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/EasyCV/modelzoo/classification/edgevit/edgexxs/ClsEvaluator_neck_top1_best.pth) |
+| EdgeVit_xs_b256_224 | [EdgeVit_xs_b256_224](https://github.com/alibaba/EasyCV/tree/master/configs/classification/imagenet/edgevit/imagenet_edgeVIT_xs_jpg.py) | 77.624   | 93.47  | 551   | 8.04    | [model](http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/EasyCV/modelzoo/classification/edgevit/edgexs/ClsEvaluator_neck_top1_best.pth) |
+| EdgeVit_s_b128_224 | [EdgeVit_s_b128_224](https://github.com/alibaba/EasyCV/tree/master/configs/classification/imagenet/edgevit/imagenet_edgeVIT_s_jpg.py) | 80.3   | 95.302  | 576   | 13.49   | [model](http://pai-vision-data-hz.oss-cn-zhangjiakou.aliyuncs.com/EasyCV/modelzoo/classification/edgevit/edges/ClsEvaluator_neck_top1_best.pth) |
 
 (ps: 通过导入官方模型得到推理结果，需要torch.__version__ >= 1.9.0，推理的输入尺寸默认为224，机器默认为V100 16G，其中gpu memory记录的是gpu peak memory)

easycv/models/backbones/__init__.py

@@ -3,6 +3,7 @@ from .benchmark_mlp import BenchMarkMLP
 from .bninception import BNInception
 from .conv_mae_vit import FastConvMAEViT
 from .conv_vitdet import ConvViTDet
+from .edgevit import EdgeVit
 from .efficientformer import EfficientFormer
 from .face_keypoint_backbone import FaceKeypointBackbone
 from .genet import PlainNet

easycv/models/backbones/edgevit.py

@@ -0,0 +1,418 @@
+# Copyright (c) Alibaba, Inc. and its affiliates.
+"""
+This model is taken from
+https://github.com/SamsungLabs/EdgeViTs
+"""
+
+from collections import OrderedDict
+from functools import partial
+
+import torch
+import torch.nn as nn
+from timm.models.layers import DropPath, to_2tuple, trunc_normal_
+
+from easycv.models.utils import ConvMlp, Mlp
+from easycv.utils.checkpoint import load_checkpoint
+from easycv.utils.logger import get_root_logger
+from ..registry import BACKBONES
+
+
+class GlobalSparseAttn(nn.Module):
+
+    def __init__(self,
+                 dim,
+                 num_heads=8,
+                 qkv_bias=False,
+                 qk_scale=None,
+                 attn_drop=0.,
+                 proj_drop=0.,
+                 sr_ratio=1):
+        super().__init__()
+        self.num_heads = num_heads
+        head_dim = dim // num_heads
+        # NOTE scale factor was wrong in my original version, can set manually to be compat with prev weights
+        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.upsample = nn.Upsample(scale_factor=sr_ratio, mode='nearest')
+        self.sr = sr_ratio
+        if self.sr > 1:
+            self.sampler = nn.AvgPool2d(1, sr_ratio)
+            kernel_size = sr_ratio
+            self.LocalProp = nn.ConvTranspose2d(
+                dim, dim, kernel_size, stride=sr_ratio, groups=dim)
+            self.norm = nn.LayerNorm(dim)
+        else:
+            self.sampler = nn.Identity()
+            self.upsample = nn.Identity()
+            self.norm = nn.Identity()
+
+    def forward(self, x, H: int, W: int):
+        B, N, C = x.shape
+        if self.sr > 1.:
+            x = x.transpose(1, 2).reshape(B, C, H, W)
+            x = self.sampler(x)
+            x = x.flatten(2).transpose(1, 2)
+
+        qkv = self.qkv(x).reshape(B, -1, 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, -1, C)
+
+        if self.sr > 1:
+            x = x.permute(0, 2, 1).reshape(B, C, int(H / self.sr),
+                                           int(W / self.sr))
+            x = self.LocalProp(x)
+            x = x.reshape(B, C, -1).permute(0, 2, 1)
+            x = self.norm(x)
+
+        x = self.proj(x)
+        x = self.proj_drop(x)
+        return x
+
+
+class LocalAgg(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.,
+                 act_layer=nn.GELU,
+                 norm_layer=nn.LayerNorm):
+        super().__init__()
+        self.pos_embed = nn.Conv2d(dim, dim, 3, padding=1, groups=dim)
+        self.norm1 = nn.BatchNorm2d(dim)
+        self.conv1 = nn.Conv2d(dim, dim, 1)
+        self.conv2 = nn.Conv2d(dim, dim, 1)
+        self.attn = nn.Conv2d(dim, dim, 5, padding=2, groups=dim)
+        self.drop_path = DropPath(
+            drop_path) if drop_path > 0. else nn.Identity()
+        self.norm2 = nn.BatchNorm2d(dim)
+        mlp_hidden_dim = int(dim * mlp_ratio)
+        self.mlp = ConvMlp(
+            in_features=dim,
+            hidden_features=mlp_hidden_dim,
+            act_layer=act_layer,
+            drop=drop)
+
+    def forward(self, x):
+        x = x + self.pos_embed(x)
+        x = x + self.drop_path(
+            self.conv2(self.attn(self.conv1(self.norm1(x)))))
+        x = x + self.drop_path(self.mlp(self.norm2(x)))
+        return x
+
+
+class SelfAttn(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.,
+                 act_layer=nn.GELU,
+                 norm_layer=nn.LayerNorm,
+                 sr_ratio=1.):
+        super().__init__()
+        self.pos_embed = nn.Conv2d(dim, dim, 3, padding=1, groups=dim)
+        self.norm1 = norm_layer(dim)
+        self.attn = GlobalSparseAttn(
+            dim,
+            num_heads=num_heads,
+            qkv_bias=qkv_bias,
+            qk_scale=qk_scale,
+            attn_drop=attn_drop,
+            proj_drop=drop,
+            sr_ratio=sr_ratio)
+        # 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)
+        # global layer_scale
+        # self.ls = layer_scale
+
+    def forward(self, x):
+        x = x + self.pos_embed(x)
+        B, N, H, W = x.shape
+        x = x.flatten(2).transpose(1, 2)
+        x = x + self.drop_path(self.attn(self.norm1(x), H, W))
+        x = x + self.drop_path(self.mlp(self.norm2(x)))
+        x = x.transpose(1, 2).reshape(B, N, H, W)
+        return x
+
+
+class LGLBlock(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.,
+                 act_layer=nn.GELU,
+                 norm_layer=nn.LayerNorm,
+                 sr_ratio=1.):
+        super().__init__()
+
+        if sr_ratio > 1:
+            self.LocalAgg = LocalAgg(dim, num_heads, mlp_ratio, qkv_bias,
+                                     qk_scale, drop, attn_drop, drop_path,
+                                     act_layer, norm_layer)
+        else:
+            self.LocalAgg = nn.Identity()
+
+        self.SelfAttn = SelfAttn(dim, num_heads, mlp_ratio, qkv_bias, qk_scale,
+                                 drop, attn_drop, drop_path, act_layer,
+                                 norm_layer, sr_ratio)
+
+    def forward(self, x):
+        x = self.LocalAgg(x)
+        x = self.SelfAttn(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.norm = nn.LayerNorm(embed_dim)
+        self.proj = nn.Conv2d(
+            in_chans, embed_dim, kernel_size=patch_size, stride=patch_size)
+
+    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 ({B}*{C}*{H}*{W}) doesn't match model ({self.img_size[0]}*{self.img_size[1]})."
+        x = self.proj(x)
+        B, C, H, W = x.shape
+        x = x.flatten(2).transpose(1, 2)
+        x = self.norm(x)
+        x = x.reshape(B, H, W, -1).permute(0, 3, 1, 2).contiguous()
+        return x
+
+
+@BACKBONES.register_module()
+class EdgeVit(nn.Module):
+    """ Vision Transformer
+    A PyTorch impl of : `An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale`  -
+        https://arxiv.org/abs/2010.11929
+    """
+
+    def __init__(self,
+                 depth=[1, 2, 3, 2],
+                 img_size=224,
+                 in_chans=3,
+                 num_classes=1000,
+                 embed_dim=[48, 96, 240, 384],
+                 head_dim=48,
+                 mlp_ratio=[4] * 4,
+                 qkv_bias=True,
+                 qk_scale=None,
+                 representation_size=None,
+                 drop_rate=0.,
+                 attn_drop_rate=0.,
+                 drop_path_rate=0.,
+                 norm_layer=partial(nn.LayerNorm, eps=1e-8),
+                 sr_ratios=[4, 2, 2, 1],
+                 pretrained=None):
+        """
+        Args:
+            depth (list): depth of each stage
+            img_size (int, tuple): input image size
+            in_chans (int): number of input channels
+            num_classes (int): number of classes for classification head
+            embed_dim (list): embedding dimension of each stage
+            head_dim (int): head dimension
+            mlp_ratio (int): ratio of mlp hidden dim to embedding dim
+            qkv_bias (bool): enable bias for qkv if True
+            qk_scale (float): override default qk scale of head_dim ** -0.5 if set
+            representation_size (Optional[int]): enable and set representation layer (pre-logits) to this value if set
+            drop_rate (float): dropout rate
+            attn_drop_rate (float): attention dropout rate
+            drop_path_rate (float): stochastic depth rate
+            norm_layer (nn.Module): normalization layer
+        """
+        super().__init__()
+        self.num_classes = num_classes
+        self.num_features = self.embed_dim = embed_dim  # num_features for consistency with other models
+        norm_layer = norm_layer or partial(nn.LayerNorm, eps=1e-6)
+
+        self.patch_embed1 = PatchEmbed(
+            img_size=img_size,
+            patch_size=4,
+            in_chans=in_chans,
+            embed_dim=embed_dim[0])
+        self.patch_embed2 = PatchEmbed(
+            img_size=img_size // 4,
+            patch_size=2,
+            in_chans=embed_dim[0],
+            embed_dim=embed_dim[1])
+        self.patch_embed3 = PatchEmbed(
+            img_size=img_size // 8,
+            patch_size=2,
+            in_chans=embed_dim[1],
+            embed_dim=embed_dim[2])
+        self.patch_embed4 = PatchEmbed(
+            img_size=img_size // 16,
+            patch_size=2,
+            in_chans=embed_dim[2],
+            embed_dim=embed_dim[3])
+
+        self.pos_drop = nn.Dropout(p=drop_rate)
+        dpr = [
+            x.item() for x in torch.linspace(0, drop_path_rate, sum(depth))
+        ]  # stochastic depth decay rule
+        num_heads = [dim // head_dim for dim in embed_dim]
+        self.blocks1 = nn.ModuleList([
+            LGLBlock(
+                dim=embed_dim[0],
+                num_heads=num_heads[0],
+                mlp_ratio=mlp_ratio[0],
+                qkv_bias=qkv_bias,
+                qk_scale=qk_scale,
+                drop=drop_rate,
+                attn_drop=attn_drop_rate,
+                drop_path=dpr[i],
+                norm_layer=norm_layer,
+                sr_ratio=sr_ratios[0]) for i in range(depth[0])
+        ])
+        self.blocks2 = nn.ModuleList([
+            LGLBlock(
+                dim=embed_dim[1],
+                num_heads=num_heads[1],
+                mlp_ratio=mlp_ratio[1],
+                qkv_bias=qkv_bias,
+                qk_scale=qk_scale,
+                drop=drop_rate,
+                attn_drop=attn_drop_rate,
+                drop_path=dpr[i + depth[0]],
+                norm_layer=norm_layer,
+                sr_ratio=sr_ratios[1]) for i in range(depth[1])
+        ])
+        self.blocks3 = nn.ModuleList([
+            LGLBlock(
+                dim=embed_dim[2],
+                num_heads=num_heads[2],
+                mlp_ratio=mlp_ratio[2],
+                qkv_bias=qkv_bias,
+                qk_scale=qk_scale,
+                drop=drop_rate,
+                attn_drop=attn_drop_rate,
+                drop_path=dpr[i + depth[0] + depth[1]],
+                norm_layer=norm_layer,
+                sr_ratio=sr_ratios[2]) for i in range(depth[2])
+        ])
+        self.blocks4 = nn.ModuleList([
+            LGLBlock(
+                dim=embed_dim[3],
+                num_heads=num_heads[3],
+                mlp_ratio=mlp_ratio[3],
+                qkv_bias=qkv_bias,
+                qk_scale=qk_scale,
+                drop=drop_rate,
+                attn_drop=attn_drop_rate,
+                drop_path=dpr[i + depth[0] + depth[1] + depth[2]],
+                norm_layer=norm_layer,
+                sr_ratio=sr_ratios[3]) for i in range(depth[3])
+        ])
+        self.norm = nn.BatchNorm2d(embed_dim[-1])
+
+        # Representation layer
+        if representation_size:
+            self.num_features = representation_size
+            self.pre_logits = nn.Sequential(
+                OrderedDict([('fc', nn.Linear(embed_dim, representation_size)),
+                             ('act', nn.Tanh())]))
+        else:
+            self.pre_logits = nn.Identity()
+
+        self.pretrained = pretrained
+        self.init_weights()
+
+    def init_weights(self, pretrained=None):
+        """Initialize the weights in backbone.
+        Args:
+            pretrained (str, optional): Path to pre-trained weights.
+                Defaults to None.
+        """
+        pretrained = pretrained or self.pretrained
+
+        def _init_weights(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)
+
+        if isinstance(pretrained, str):
+            self.apply(_init_weights)
+            logger = get_root_logger()
+            load_checkpoint(self, pretrained, strict=False, logger=logger)
+        elif pretrained is None:
+            self.apply(_init_weights)
+        else:
+            raise TypeError('pretrained must be a str or None')
+
+    @torch.jit.ignore
+    def no_weight_decay(self):
+        return {'pos_embed', 'cls_token'}
+
+    def forward_features(self, x):
+        x = self.patch_embed1(x)
+        x = self.pos_drop(x)
+        for blk in self.blocks1:
+            x = blk(x)
+        x = self.patch_embed2(x)
+        for blk in self.blocks2:
+            x = blk(x)
+        x = self.patch_embed3(x)
+        for blk in self.blocks3:
+            x = blk(x)
+        x = self.patch_embed4(x)
+        for blk in self.blocks4:
+            x = blk(x)
+        x = self.norm(x)
+        x = self.pre_logits(x)
+        return x
+
+    def forward(self, x):
+        x = self.forward_features(x)
+        return [x]

easycv/models/utils/__init__.py

@@ -18,7 +18,7 @@ from .scale import Scale
 # from .weight_init import (bias_init_with_prob, kaiming_init, normal_init,
 #                          uniform_init, xavier_init)
 from .sobel import Sobel
-from .transformer import (MLP, DropPath, Mlp, TransformerEncoder,
+from .transformer import (MLP, ConvMlp, DropPath, Mlp, TransformerEncoder,
                           TransformerEncoderLayer, _get_activation_fn,
                           _get_clones)
 

easycv/models/utils/transformer.py

@@ -66,6 +66,31 @@ class Mlp(nn.Module):
         return x
 
 
+class ConvMlp(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.Conv2d(in_features, hidden_features, 1)
+        self.act = act_layer()
+        self.fc2 = nn.Conv2d(hidden_features, out_features, 1)
+        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
+
+
 def drop_path(x, drop_prob: float = 0., training: bool = False):
     if drop_prob == 0. or not training:
         return x

tests/models/backbones/test_edgevit.py

@@ -0,0 +1,36 @@
+# Copyright (c) Alibaba, Inc. and its affiliates.
+import unittest
+
+import torch
+
+from easycv.models.backbones import EdgeVit
+
+
+class EdgeVitTest(unittest.TestCase):
+
+    def setUp(self):
+        print(('Testing %s.%s' % (type(self).__name__, self._testMethodName)))
+
+    def test_vitdet(self):
+        model = EdgeVit(
+            img_size=224,
+            depth=[1, 1, 3, 2],
+            embed_dim=[36, 72, 144, 288],
+            head_dim=36,
+            mlp_ratio=[4] * 4,
+            qkv_bias=True,
+            num_classes=1000,
+            drop_path_rate=0.1,
+            sr_ratios=[4, 2, 2, 1],
+        )
+
+        model.init_weights()
+        model.train()
+        imgs = torch.rand(36, 3, 224, 224)
+        feat = model(imgs)
+        self.assertEqual(len(feat), 1)
+        self.assertEqual(feat[0].shape, torch.Size([36, 288, 7, 7]))
+
+
+if __name__ == '__main__':
+    unittest.main()