File size: 5,075 Bytes
eb0678a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0975142
eb0678a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
#    This file may have been modified by Flash-VStream Authors (Flash-VStream Modifications”). All Flash-VStream Modifications are Copyright 2024 Flash-VStream Authors. 
# ------------------------------------------------------------------------
# Based on https://github.com/haotian-liu/LLaVA. Below is the original copyright:
#    Copyright 2023 Haotian Liu
#
#    Licensed under the Apache License, Version 2.0 (the "License");
#    you may not use this file except in compliance with the License.
#    You may obtain a copy of the License at
#
#        http://www.apache.org/licenses/LICENSE-2.0
#
#    Unless required by applicable law or agreed to in writing, software
#    distributed under the License is distributed on an "AS IS" BASIS,
#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#    See the License for the specific language governing permissions and
#    limitations under the License.

import torch
import torch.nn as nn
import torch.nn.functional as F
from typing import List, Optional, Tuple, Union
from transformers import AutoConfig, AutoModelForCausalLM, \
                         LlamaConfig, LlamaModel, LlamaForCausalLM
from transformers.modeling_outputs import CausalLMOutputWithPast
from flash_vstream.model.vstream_arch import VStreamMetaModel, VStreamMetaForCausalLM


class VStreamConfig(LlamaConfig):
    model_type = "vstream"


class VStreamLlamaModel(VStreamMetaModel, LlamaModel):
    config_class = VStreamConfig

    def __init__(self, config: LlamaConfig):
        super(VStreamLlamaModel, self).__init__(config)


class VStreamLlamaForCausalLM(VStreamMetaForCausalLM, LlamaForCausalLM):
    config_class = VStreamConfig

    def __init__(self, config):
        super(VStreamLlamaForCausalLM, self).__init__(config)
        self.model = VStreamLlamaModel(config)
        self.pretraining_tp = config.pretraining_tp
        self.vocab_size = config.vocab_size
        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)

        # Initialize weights and apply final processing
        self.post_init()

    def get_model(self):
        return self.model
    
    def forward(
        self,
        input_ids: torch.LongTensor = None,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[List[torch.FloatTensor]] = None,
        inputs_embeds: Optional[torch.FloatTensor] = None,
        labels: Optional[torch.LongTensor] = None,
        use_cache: Optional[bool] = True,
        output_attentions: Optional[bool] = None,
        output_hidden_states: Optional[bool] = None,
        images: Optional[torch.FloatTensor] = None,
        features: Optional[torch.FloatTensor] = None,
        return_dict: Optional[bool] = None,
        cache_position=None,
    ) -> Union[Tuple, CausalLMOutputWithPast]:
        if inputs_embeds is None:
            if self.use_video_streaming_mode:
                (
                    input_ids,
                    position_ids,
                    attention_mask,
                    past_key_values,
                    inputs_embeds,
                    labels
                ) = self.prepare_inputs_labels_for_multimodal_streaming(
                    input_ids,
                    position_ids,
                    attention_mask,
                    past_key_values,
                    labels,
                )
            else:
                (
                    input_ids,
                    position_ids,
                    attention_mask,
                    past_key_values,
                    inputs_embeds,
                    labels
                ) = self.prepare_inputs_labels_for_multimodal(
                    input_ids,
                    position_ids,
                    attention_mask,
                    past_key_values,
                    labels,
                    images,
                    features,
                )
        return super().forward(
            input_ids=input_ids,
            attention_mask=attention_mask,
            position_ids=position_ids,
            past_key_values=past_key_values,
            inputs_embeds=inputs_embeds,
            labels=labels,
            use_cache=use_cache,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_dict=return_dict
        )

    def prepare_inputs_for_generation(self, input_ids, past_key_values=None, inputs_embeds=None, **kwargs):
        images = kwargs.pop("images", None)
        features = kwargs.pop("features", None)
        _inputs = super().prepare_inputs_for_generation(
            input_ids, past_key_values=past_key_values, inputs_embeds=inputs_embeds, **kwargs
        )
        if images is not None:
            _inputs['images'] = images
        if features is not None:
            _inputs['features'] = features
        return _inputs

AutoConfig.register("vstream", VStreamConfig)
AutoModelForCausalLM.register(VStreamConfig, VStreamLlamaForCausalLM)