File size: 19,803 Bytes
26c4860
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import logging

import torch
from torch import nn
import torch.nn.functional as F

from modules.until_module import PreTrainedModel, AllGather, CrossEn
from modules.module_cross import CrossModel, CrossConfig, Transformer as TransformerClip

from modules.module_clip import CLIP, convert_weights
from torch.nn.utils.rnn import pad_packed_sequence, pack_padded_sequence

logger = logging.getLogger(__name__)
allgather = AllGather.apply

class CLIP4ClipPreTrainedModel(PreTrainedModel, nn.Module):
    """ An abstract class to handle weights initialization and
        a simple interface for dowloading and loading pretrained models.
    """
    def __init__(self, cross_config, *inputs, **kwargs):
        super(CLIP4ClipPreTrainedModel, self).__init__(cross_config)
        self.cross_config = cross_config
        self.clip = None
        self.clip_tag = None   #新增:tag编码器

        self.cross = None

    @classmethod
    
    def from_pretrained(cls, cross_model_name, state_dict=None, cache_dir=None, type_vocab_size=2, *inputs, **kwargs):

        task_config = None
        if "task_config" in kwargs.keys():
            task_config = kwargs["task_config"]
            if not hasattr(task_config, "local_rank"):
                task_config.__dict__["local_rank"] = 0
            elif task_config.local_rank == -1:
                task_config.local_rank = 0

        if state_dict is None: state_dict = {}
        pretrained_clip_name = "ViT-B/32"    #默认使用"ViT-B/32"作为预训练的CLIP模型名称,但如果task_config中有指定,则使用指定的名称。
        if hasattr(task_config, 'pretrained_clip_name'):
            pretrained_clip_name = task_config.pretrained_clip_name

        #加载clip预训练权重
        clip_state_dict = CLIP.get_config(pretrained_clip_name=pretrained_clip_name)


        # 为两个文本编码器准备权重
        for key, val in clip_state_dict.items():
            # 主文本编码器权重
            new_key = "clip." + key
            if new_key not in state_dict:
                state_dict[new_key] = val.clone()
            # 标签文本编码器权重
            new_key = "clip_tag." + key
            if new_key not in state_dict:
                state_dict[new_key] = val.clone()

        cross_config, _ = CrossConfig.get_config(cross_model_name, cache_dir, type_vocab_size, state_dict=None, task_config=task_config)

        model = cls(cross_config, clip_state_dict, *inputs, **kwargs)

        ## ===> Initialization trick [HARD CODE]
        if model.linear_patch == "3d":
            contain_conv2 = False
            for key in state_dict.keys():
                if key.find("visual.conv2.weight") > -1:
                    contain_conv2 = True
                    break
            if contain_conv2 is False and hasattr(model.clip.visual, "conv2"):
                cp_weight = state_dict["clip.visual.conv1.weight"].clone()
                kernel_size = model.clip.visual.conv2.weight.size(2)
                conv2_size = model.clip.visual.conv2.weight.size()
                conv2_size = list(conv2_size)

                left_conv2_size = conv2_size.copy()
                right_conv2_size = conv2_size.copy()
                left_conv2_size[2] = (kernel_size - 1) // 2
                right_conv2_size[2] = kernel_size - 1 - left_conv2_size[2]

                left_zeros, right_zeros = None, None
                if left_conv2_size[2] > 0:
                    left_zeros = torch.zeros(*tuple(left_conv2_size), dtype=cp_weight.dtype, device=cp_weight.device)
                if right_conv2_size[2] > 0:
                    right_zeros = torch.zeros(*tuple(right_conv2_size), dtype=cp_weight.dtype, device=cp_weight.device)

                cat_list = []
                if left_zeros != None: cat_list.append(left_zeros)
                cat_list.append(cp_weight.unsqueeze(2))
                if right_zeros != None: cat_list.append(right_zeros)
                cp_weight = torch.cat(cat_list, dim=2)

                state_dict["clip.visual.conv2.weight"] = cp_weight
                state_dict["clip_tag.visual.conv2.weight"] = cp_weight.clone()  # 同时更新两个编码器

        ## <=== End of initialization trick

        if state_dict is not None:
            model = cls.init_preweight(model, state_dict, task_config=task_config)

        return model

def show_log(task_config, info):
    if task_config is None or task_config.local_rank == 0:
        logger.warning(info)

def update_attr(target_name, target_config, target_attr_name, source_config, source_attr_name, default_value=None):
    if hasattr(source_config, source_attr_name):
        if default_value is None or getattr(source_config, source_attr_name) != default_value:
            setattr(target_config, target_attr_name, getattr(source_config, source_attr_name))
            show_log(source_config, "Set {}.{}: {}.".format(target_name,
                                                            target_attr_name, getattr(target_config, target_attr_name)))
    return target_config

def check_attr(target_name, task_config):
    return hasattr(task_config, target_name) and task_config.__dict__[target_name]


class DynamicFusion(nn.Module):
    def __init__(self, hidden_size, text_bias=0.7):
        super().__init__()
        self.text_proj = nn.Linear(hidden_size, hidden_size)
        self.tag_proj = nn.Linear(hidden_size, hidden_size)
        
        # 注意力计算层
        self.attention = nn.Sequential(
            nn.Linear(hidden_size, hidden_size),  # 修改为 hidden_size
            nn.Tanh(),
            nn.Linear(hidden_size, 1)
        )
        
        # 可学习的偏置参数
        self.text_bias = nn.Parameter(torch.tensor([text_bias]))
        self.tag_bias = nn.Parameter(torch.tensor([1 - text_bias]))
    
    def forward(self, text_feat, tag_feat):
        # 特征投影
        text_proj = self.text_proj(text_feat.squeeze(1))  # 去掉多余的维度 [16,512]
        tag_proj = self.tag_proj(tag_feat.squeeze(1))    # 去掉多余的维度 [16,512]
        
        # 计算注意力权重
        combined = torch.stack([text_proj, tag_proj], dim=1)  # [16, 2, 512]
        batch_size = combined.size(0)
        
        # 展平用于注意力计算
        flattened = combined.view(-1, combined.size(-1))  # [32, 512]
        raw_weights = self.attention(flattened)  # [32, 1]
        raw_weights = raw_weights.view(batch_size, 2, 1)  # [16, 2, 1]
        
        # 加入偏置
        biased_weights = raw_weights + torch.cat([self.text_bias, self.tag_bias], dim=0).view(1, 2, 1)
        attn_weights = F.softmax(biased_weights, dim=1)

        # 加权融合
        fused_feat = attn_weights[:, 0] * text_proj + attn_weights[:, 1] * tag_proj
        return fused_feat.unsqueeze(1)  # 恢复为 [16, 1, 512]

class CLIP4Clip(CLIP4ClipPreTrainedModel):
    def __init__(self, cross_config, clip_state_dict, task_config):
        super(CLIP4Clip, self).__init__(cross_config)

        self.task_config = task_config
        self.ignore_video_index = -1

        # 1.在 CLIP4Clip 的 __init__ 方法中调整
        cross_config.max_position_embeddings = self.task_config.max_words + self.task_config.max_frames

        # 2.重新初始化位置嵌入
        self.cross = CrossModel(cross_config)  # 重新加载交叉编码器

        #assert self.task_config.max_words + self.task_config.max_frames <= cross_config.max_position_embeddings

        self._stage_one = True
        self._stage_two = False

        show_log(task_config, "Stage-One:{}, Stage-Two:{}".format(self._stage_one, self._stage_two))

        self.loose_type = False
        if self._stage_one and check_attr('loose_type', self.task_config):
            self.loose_type = True
            show_log(task_config, "Test retrieval by loose type.")

        # CLIP Encoders: From OpenAI: CLIP [https://github.com/openai/CLIP] ===>
        vit = "visual.proj" in clip_state_dict
        assert vit
        if vit:
            vision_width = clip_state_dict["visual.conv1.weight"].shape[0]
            vision_layers = len(
                [k for k in clip_state_dict.keys() if k.startswith("visual.") and k.endswith(".attn.in_proj_weight")])
            vision_patch_size = clip_state_dict["visual.conv1.weight"].shape[-1]
            grid_size = round((clip_state_dict["visual.positional_embedding"].shape[0] - 1) ** 0.5)
            image_resolution = vision_patch_size * grid_size
        else:
            counts: list = [len(set(k.split(".")[2] for k in clip_state_dict if k.startswith(f"visual.layer{b}"))) for b in
                            [1, 2, 3, 4]]
            vision_layers = tuple(counts)
            vision_width = clip_state_dict["visual.layer1.0.conv1.weight"].shape[0]
            output_width = round((clip_state_dict["visual.attnpool.positional_embedding"].shape[0] - 1) ** 0.5)
            vision_patch_size = None
            assert output_width ** 2 + 1 == clip_state_dict["visual.attnpool.positional_embedding"].shape[0]
            image_resolution = output_width * 32

        embed_dim = clip_state_dict["text_projection"].shape[1]  #文本编码器参数提取
        context_length = clip_state_dict["positional_embedding"].shape[0]
        vocab_size = clip_state_dict["token_embedding.weight"].shape[0]
        transformer_width = clip_state_dict["ln_final.weight"].shape[0]
        transformer_heads = transformer_width // 64
        transformer_layers = len(set(k.split(".")[2] for k in clip_state_dict if k.startswith(f"transformer.resblocks")))

        show_log(task_config, "\t embed_dim: {}".format(embed_dim))
        show_log(task_config, "\t image_resolution: {}".format(image_resolution))
        show_log(task_config, "\t vision_layers: {}".format(vision_layers))
        show_log(task_config, "\t vision_width: {}".format(vision_width))
        show_log(task_config, "\t vision_patch_size: {}".format(vision_patch_size))
        show_log(task_config, "\t context_length: {}".format(context_length))
        show_log(task_config, "\t vocab_size: {}".format(vocab_size))
        show_log(task_config, "\t transformer_width: {}".format(transformer_width))
        show_log(task_config, "\t transformer_heads: {}".format(transformer_heads))
        show_log(task_config, "\t transformer_layers: {}".format(transformer_layers))

        self.linear_patch = '2d'
        if hasattr(task_config, "linear_patch"):
            self.linear_patch = task_config.linear_patch
            show_log(task_config, "\t\t linear_patch: {}".format(self.linear_patch))

        cut_top_layer = 0
        show_log(task_config, "\t cut_top_layer: {}".format(cut_top_layer))

        # 初始化编码器
        self.clip = CLIP(
            embed_dim,
            image_resolution, vision_layers-cut_top_layer, vision_width, vision_patch_size,
            context_length, vocab_size, transformer_width, transformer_heads, transformer_layers-cut_top_layer,
            linear_patch=self.linear_patch
        ).float()

        # 标签文本编码器 - 使用不同的初始化
        self.clip_tag = CLIP(
            embed_dim,
            image_resolution, vision_layers-cut_top_layer, vision_width, vision_patch_size,
            context_length, vocab_size, transformer_width, transformer_heads, transformer_layers-cut_top_layer,
            linear_patch=self.linear_patch
        ).float()
        
        # 共享视觉编码器权重
        self.clip_tag.visual.load_state_dict(self.clip.visual.state_dict())

        # 转换权重
        convert_weights(self.clip_tag)
        convert_weights(self.clip)

        for key in ["input_resolution", "context_length", "vocab_size"]:
            if key in clip_state_dict:
                del clip_state_dict[key]

        # <=== End of CLIP Encoders

        self.sim_header = 'meanP'
        if hasattr(task_config, "sim_header"):
            self.sim_header = task_config.sim_header
            show_log(task_config, "\t sim_header: {}".format(self.sim_header))

        cross_config.max_position_embeddings = context_length
        if self.loose_type is False:
            # Cross Encoder ===>
            cross_config = update_attr("cross_config", cross_config, "num_hidden_layers", self.task_config, "cross_num_hidden_layers")
            self.cross = CrossModel(cross_config)
            # <=== End of Cross Encoder
            self.similarity_dense = nn.Linear(cross_config.hidden_size, 1)

        self.loss_fct = CrossEn()

        self.apply(self.init_weights)

        self.dynamic_fusion = DynamicFusion(cross_config.hidden_size)    ######

    def forward(self, input_ids, token_type_ids, attention_mask, input_ids_tag, token_type_ids_tag, attention_mask_tag, video, video_mask=None):
        input_ids = input_ids.view(-1, input_ids.shape[-1])
        token_type_ids = token_type_ids.view(-1, token_type_ids.shape[-1])
        attention_mask = attention_mask.view(-1, attention_mask.shape[-1])

        input_ids_tag = input_ids_tag.view(-1, input_ids_tag.shape[-1])
        token_type_ids_tag = token_type_ids_tag.view(-1, token_type_ids_tag.shape[-1])
        attention_mask_tag = attention_mask_tag.view(-1, attention_mask_tag.shape[-1])

        video_mask = video_mask.view(-1, video_mask.shape[-1])

        # T x 3 x H x W
        video = torch.as_tensor(video).float()
        b, pair, bs, ts, channel, h, w = video.shape
        video = video.view(b * pair * bs * ts, channel, h, w)
        video_frame = bs * ts

        sequence_output, sequence_output_tag, visual_output = self.get_sequence_visual_output(input_ids, token_type_ids, attention_mask,input_ids_tag, token_type_ids_tag,attention_mask_tag,video, video_mask, shaped=True, video_frame=video_frame)   #3.model 类的自定义方法调用栈

        fused_features = self.dynamic_fusion(sequence_output, sequence_output_tag)


        if self.training:
            loss = 0.
            sim_matrix, *_tmp = self.get_similarity_logits(fused_features, visual_output, attention_mask, video_mask, shaped=True, loose_type=self.loose_type)
            sim_loss1 = self.loss_fct(sim_matrix)
            sim_loss2 = self.loss_fct(sim_matrix.T)
            sim_loss = (sim_loss1 + sim_loss2) / 2
            loss += sim_loss

            return loss
        else:
            return None

    def get_sequence_output(self, input_ids, token_type_ids, attention_mask, shaped=False):
        if shaped is False:
            input_ids = input_ids.view(-1, input_ids.shape[-1])
            token_type_ids = token_type_ids.view(-1, token_type_ids.shape[-1])
            attention_mask = attention_mask.view(-1, attention_mask.shape[-1])

        bs_pair = input_ids.size(0)
        sequence_hidden = self.clip.encode_text(input_ids).float()
        sequence_hidden = sequence_hidden.view(bs_pair, -1, sequence_hidden.size(-1))

        return sequence_hidden


    def get_sequence_output_tag(self, input_ids, token_type_ids, attention_mask, shaped=False):
        if shaped is False:
            input_ids = input_ids.view(-1, input_ids.shape[-1])
            token_type_ids = token_type_ids.view(-1, token_type_ids.shape[-1])
            attention_mask = attention_mask.view(-1, attention_mask.shape[-1])

        bs_pair = input_ids.size(0)
        sequence_hidden_tag = self.clip_tag.encode_text(input_ids).float()
        sequence_hidden_tag = sequence_hidden_tag.view(bs_pair, -1, sequence_hidden_tag.size(-1))

        return sequence_hidden_tag

    def get_visual_output(self, video, video_mask, shaped=False, video_frame=-1):
        if shaped is False:
            video_mask = video_mask.view(-1, video_mask.shape[-1])
            video = torch.as_tensor(video).float()
            b, pair, bs, ts, channel, h, w = video.shape
            video = video.view(b * pair * bs * ts, channel, h, w)
            video_frame = bs * ts

        bs_pair = video_mask.size(0)
        visual_hidden = self.clip.encode_image(video, video_frame=video_frame).float()
        visual_hidden = visual_hidden.view(bs_pair, -1, visual_hidden.size(-1))

        return visual_hidden   #获取视频特征

    def get_sequence_visual_output(self, input_ids, token_type_ids, attention_mask, input_ids_tag, token_type_ids_tag, attention_mask_tag, video, video_mask, shaped=False, video_frame=-1):
        if shaped is False:
            input_ids = input_ids.view(-1, input_ids.shape[-1])
            token_type_ids = token_type_ids.view(-1, token_type_ids.shape[-1])
            attention_mask = attention_mask.view(-1, attention_mask.shape[-1])

            input_ids_tag = input_ids_tag.view(-1, input_ids_tag.shape[-1])
            token_type_ids_tag = token_type_ids_tag.view(-1, token_type_ids_tag.shape[-1])
            attention_mask_tag = attention_mask_tag.view(-1, attention_mask_tag.shape[-1])


            video_mask = video_mask.view(-1, video_mask.shape[-1])

            video = torch.as_tensor(video).float()
            b, pair, bs, ts, channel, h, w = video.shape
            video = video.view(b * pair * bs * ts, channel, h, w)
            video_frame = bs * ts

        sequence_output = self.get_sequence_output(input_ids, token_type_ids, attention_mask, shaped=True)
        sequence_output_tag = self.get_sequence_output_tag(input_ids_tag, token_type_ids_tag, attention_mask_tag, shaped=True)
        visual_output = self.get_visual_output(video, video_mask, shaped=True, video_frame=video_frame)

        return sequence_output, sequence_output_tag, visual_output   #同时获取文本和视频的特征表示

    def _mean_pooling_for_similarity_visual(self, visual_output, video_mask,):   #对视频特征序列进行掩码加权平均池化 
        video_mask_un = video_mask.to(dtype=torch.float).unsqueeze(-1)
        visual_output = visual_output * video_mask_un
        video_mask_un_sum = torch.sum(video_mask_un, dim=1, dtype=torch.float)
        video_mask_un_sum[video_mask_un_sum == 0.] = 1.
        video_out = torch.sum(visual_output, dim=1) / video_mask_un_sum
        return video_out


    def _loose_similarity(self, fused_features, visual_output, attention_mask, video_mask, sim_header="meanP"):
        fused_features, visual_output = fused_features.contiguous(), visual_output.contiguous()

        if self.training and torch.distributed.is_initialized():
            visual_output = allgather(visual_output, self.task_config)
            video_mask = allgather(video_mask, self.task_config)
            fused_features = allgather(fused_features, self.task_config)
            torch.distributed.barrier()

        visual_output = visual_output / visual_output.norm(dim=-1, keepdim=True)
        visual_output = self._mean_pooling_for_similarity_visual(visual_output, video_mask)
        visual_output = visual_output / visual_output.norm(dim=-1, keepdim=True)

        fused_features = fused_features.squeeze(1)
        fused_features = fused_features / fused_features.norm(dim=-1, keepdim=True)

        logit_scale = self.clip.logit_scale.exp()
        retrieve_logits = logit_scale * torch.matmul(fused_features, visual_output.t())
        return retrieve_logits

    def get_similarity_logits(self, fused_features, visual_output, attention_mask, video_mask, shaped=False, loose_type=False):
        if shaped is False:
            attention_mask = attention_mask.view(-1, attention_mask.shape[-1])
            video_mask = video_mask.view(-1, video_mask.shape[-1])

        contrastive_direction = ()
        retrieve_logits = self._loose_similarity(fused_features, visual_output, attention_mask, video_mask, sim_header=self.sim_header)
        return retrieve_logits, contrastive_direction