File size: 10,495 Bytes
2d9a728
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import logging

import torch
from torch import nn
import numpy as np
from PIL import Image
import torchvision.transforms as transforms
from torchvision.transforms import InterpolationMode

from .backbones.internvideo2 import InternVideo2, LLaMA, Tokenizer
from .criterions import VTC_VTM_Loss

logger = logging.getLogger(__name__)


class InternVideo2_CLIP(nn.Module):
    def __init__(self, config, tokenizer=None, is_pretrain=True):
        super().__init__()

        self.config = config
        self.tokenizer = tokenizer
        self.is_pretrain = is_pretrain

        # create modules.
        if tokenizer is None:
            self.tokenizer = Tokenizer(config.model.tokenizer_path)
        self.vision_encoder = self.build_vision_encoder()
        self.text_encoder = self.build_text_encoder()
        # adopt 1 / 100. as in ViCLIP
        self.temp = nn.parameter.Parameter(torch.ones([]) * config.model.temp)
        self.temp_min = config.model.temp_min
        
        # freeze model
        if self.config.model.freeze_vision:
            for name, p in self.vision_encoder.named_parameters():
                if self.config.model.open_vision_clip_projector and name.startswith('clip_projector'):
                    logger.info(f"Unfreeze {name}")
                else:
                    logger.info(f"Freeze {name}")
                    p.requires_grad = False
        if self.config.model.freeze_text:
            for name, p in self.text_encoder.named_parameters():
                if self.config.model.open_text_projection and name.startswith('text_projection'):
                    logger.info(f"Unfreeze {name}")
                elif self.config.model.open_text_lora and 'lora' in name:
                    logger.info(f"Unfreeze {name}")
                else:
                    logger.info(f"Freeze {name}")
                    p.requires_grad = False

        img_size = self.config.model.vision_encoder.img_size
        self.transform = transforms.Compose(
            [
                transforms.Resize(
                    (img_size, img_size),
                    interpolation=InterpolationMode.BICUBIC,
                ),
                transforms.Lambda(lambda x: x.float().div(255.0)),
                transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225)),
            ]
        )
        
        # load pretrained models
        self.load_checkpoint(
            config.model.vision_ckpt_path, config.model.text_ckpt_path, 
            config.model.get("extra_ckpt_path", None)
        )
        
        # criterions
        self.clip_loss = VTC_VTM_Loss(False)

    def no_weight_decay(self):
        ret = {"temp"}
        ret.update(
            {"vision_encoder." + k for k in self.vision_encoder.no_weight_decay()}
        )
        # no weight decay for LLM if training
        ret.update(
            {"text_encoder." + k for k, _ in self.text_encoder.named_parameters()}
        )

        return ret
    
    @torch.no_grad()
    def clip_contrastive_temperature(self):
        """Seems only used during pre-training"""
        self.temp.clamp_(min=self.temp_min)

    def forward(self, image, text, idx):
        """forward and calculate loss.

        Args:
            image (torch.Tensor): The input images. Shape: [B,T,C,H,W].
            text (dict): TODO
            idx (torch.Tensor): TODO

        Returns: TODO

        """
        self.clip_contrastive_temperature()
        vision_embeds = self.encode_vision(image)
        text_embeds = self.encode_text(text)

        # VTC loss
        loss_vtc = self.clip_loss.vtc_loss(
            vision_embeds, text_embeds, idx, self.temp, all_gather=True
        )

        return dict(
            loss_vtc=loss_vtc,
        )

    def encode_vision(self, image, test=False):
        """encode image / videos as features.

        Args:
            image (torch.Tensor): The input images.
            test (bool): Whether testing.

        Returns: tuple.
            - vision_embeds (torch.Tensor): The features of all patches. Shape: [B,C].

        """
        T = image.shape[1]
        use_image = True if T == 1 else False
        image = image.permute(0, 2, 1, 3, 4) # [B,T,C,H,W] -> [B,C,T,H,W]

        vision_embeds = self.vision_encoder(image, use_image=use_image)
        return vision_embeds

    def encode_text(self, text):
        """encode text.
        Args:
            text (dict): The output of huggingface's `PreTrainedTokenizer`. contains keys:
                - input_ids (torch.Tensor): Token ids to be fed to a model. Shape: [B,L].
                - attention_mask (torch.Tensor): The mask indicate padded tokens. Shape: [B,L]. 0 is padded token.
                - other keys refer to "https://huggingface.co/docs/transformers/v4.21.2/en/main_classes/tokenizer#transformers.PreTrainedTokenizer.__call__".
        Returns: tuple.
            - text_embeds (torch.Tensor): The features of all tokens. Shape: [B,C].

        """
        text_embeds = self.text_encoder(text)
        return text_embeds

    def build_vision_encoder(self):
        """build vision encoder
        Returns: (vision_encoder, vision_layernorm). Each is a `nn.Module`.

        """
        vision_encoder = InternVideo2(
            in_chans=self.config.model.vision_encoder.in_chans,
            patch_size=self.config.model.vision_encoder.patch_size,
            img_size=self.config.model.vision_encoder.img_size,
            qkv_bias=self.config.model.vision_encoder.qkv_bias,
            drop_path_rate=self.config.model.vision_encoder.drop_path_rate,
            head_drop_path_rate=self.config.model.vision_encoder.head_drop_path_rate,
            embed_dim=self.config.model.vision_encoder.embed_dim,
            num_heads=self.config.model.vision_encoder.num_heads,
            mlp_ratio=self.config.model.vision_encoder.mlp_ratio,
            init_values=self.config.model.vision_encoder.init_values,
            qk_normalization=self.config.model.vision_encoder.qk_normalization,
            depth=self.config.model.vision_encoder.depth,
            use_flash_attn=self.config.model.vision_encoder.use_flash_attn,
            use_fused_rmsnorm=self.config.model.vision_encoder.use_fused_rmsnorm,
            use_fused_mlp=self.config.model.vision_encoder.use_fused_mlp,
            fused_mlp_heuristic=self.config.model.vision_encoder.fused_mlp_heuristic,
            attn_pool_num_heads=self.config.model.vision_encoder.attn_pool_num_heads,
            clip_embed_dim=self.config.model.vision_encoder.clip_embed_dim,
            layerscale_no_force_fp32=self.config.model.vision_encoder.layerscale_no_force_fp32,
            num_frames=self.config.model.vision_encoder.num_frames,
            tubelet_size=self.config.model.vision_encoder.tubelet_size,
            sep_pos_embed=self.config.model.vision_encoder.sep_pos_embed,
            use_checkpoint=self.config.model.vision_encoder.use_checkpoint,
            checkpoint_num=self.config.model.vision_encoder.checkpoint_num,
        )
        return vision_encoder

    def build_text_encoder(self):
        """build text_encoder and possiblly video-to-text multimodal fusion encoder.
        Returns: nn.Module. The text encoder

        """
        text_encoder = LLaMA(
            use_flash_attn=self.config.model.text_encoder.use_flash_attn,
            transformer_width=self.config.model.text_encoder.transformer_width,
            llama_path=self.config.model.text_encoder.llama_path,
            use_lora=self.config.model.text_encoder.use_lora,
        )

        return text_encoder

    def load_checkpoint(self, vision_ckpt_path=None, text_ckpt_path=None, extra_ckpt_path=None):
        assert vision_ckpt_path is not None, "No vision_encoder checkpoint"
        assert text_ckpt_path is not None, "No text_encoder checkpoint"

        new_ckpt = {}

        # load vision_encoder
        logger.info(f"Load vision_encoder checkpoint from {vision_ckpt_path}")
        vision_ckpt = torch.load(vision_ckpt_path, map_location='cpu')
        if 'module' in vision_ckpt.keys():
            vision_ckpt = vision_ckpt['module']
        elif 'model' in vision_ckpt.keys():
            vision_ckpt = vision_ckpt['model']
        if self.config.model.get('load_vision_ckpt_from_internvideo2_stage2', False):
            from .backbones.internvideo2.pos_embed import interpolate_pos_embed
            orig_t_size = self.config.model.get('vision_ckpt_t_size', 4)
            interpolate_pos_embed(vision_ckpt, self.vision_encoder, orig_t_size=orig_t_size) # 4 for InternVideo2 stage2
            for k, v in vision_ckpt.items():
                if k.startswith('vision_encoder.'):
                    if 'clip_decoder' in k or 'final_clip_decoder' in k:
                        continue
                    elif 'clip_pos_embed' in k or 'clip_img_pos_embed' in k or 'img_pos_embed' in k :
                        continue
                    else:
                        new_ckpt[k] = v
                else:
                    continue
        else:
            for k, v in vision_ckpt.items():
                if k.startswith('clip_decoder.') or k.startswith('mae_decoder.') or k.startswith('final_clip_decoder.'):
                    continue
                elif k in ['clip_pos_embed', 'mae_pos_embed']:
                    continue
                else:
                    new_k = 'vision_encoder.' + k
                    new_ckpt[new_k] = v

        # load text_encoder
        logger.info(f"Load text_encoder checkpoint from {text_ckpt_path}")
        test_ckpt = torch.load(text_ckpt_path, map_location='cpu')
        if 'module' in test_ckpt.keys():
            test_ckpt = test_ckpt['module']
        for k, v in test_ckpt.items():
            if k.startswith('transformer.') or k == 'text_projection':
                new_k = "text_encoder." + k
            else:
                continue
            new_ckpt[new_k] = v

        # load extra checkpoint
        # often when post-pretrain after previous pretraining, thus the keys are same
        if extra_ckpt_path is not None:
            logger.info(f"Load extra checkpoint from {extra_ckpt_path}")
            extra_ckpt = torch.load(extra_ckpt_path, map_location='cpu')
            if 'module' in extra_ckpt.keys():
                extra_ckpt = extra_ckpt['module']
            for k, v in extra_ckpt.items():
                new_ckpt[k] = v
        
        msg = self.load_state_dict(new_ckpt, strict=False)
        logger.info(msg)