Skip to content

GitLab

"tests/models/marian/test_tokenization_marian.py" did not exist on "71356034232cfce20d8f24df1213944665e7fe08"
Switch branch/tag
Find file Normal viewHistoryPermalink
qwen2_5_vl.py 4.99 KB
Newer Older
chenych's avatar
First commit.  
chenych committed
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 
from typing import Optional

import torch
from transformers.models.qwen2_5_vl.processing_qwen2_5_vl import Qwen2_5_VLProcessor


def get_rope_index(
    processor: Qwen2_5_VLProcessor,
    input_ids: torch.Tensor,
    image_grid_thw: Optional[torch.Tensor] = None,
    video_grid_thw: Optional[torch.Tensor] = None,
    second_per_grid_ts: Optional[torch.Tensor] = None,
    attention_mask: Optional[torch.Tensor] = None,
) -> torch.Tensor:
    """
    Gets the position ids for Qwen2-VL, it should be generated before sharding the sequence.
    The batch dim has been removed and the input_ids should be a 1D tensor representing a single example.
    https://github.com/huggingface/transformers/blob/v4.49.0/src/transformers/models/qwen2_5_vl/modeling_qwen2_5_vl.py#L1546
    """
    spatial_merge_size = processor.image_processor.merge_size
    tokens_per_second = 2
    image_token_id = processor.tokenizer.convert_tokens_to_ids("<|image_pad|>")
    video_token_id = processor.tokenizer.convert_tokens_to_ids("<|video_pad|>")
    vision_start_token_id = processor.tokenizer.convert_tokens_to_ids("<|vision_start|>")
    if input_ids is not None and (image_grid_thw is not None or video_grid_thw is not None):
        if attention_mask is None:
            attention_mask = torch.ones_like(input_ids)

        position_ids = torch.ones(3, input_ids.size(0), dtype=input_ids.dtype, device=input_ids.device)  # (3, seqlen)
        image_index, video_index = 0, 0
        input_ids = input_ids[attention_mask == 1]
        image_nums, video_nums = 0, 0
        vision_start_indices = torch.argwhere(input_ids == vision_start_token_id)
        vision_tokens = input_ids[vision_start_indices + 1]
        image_nums = (vision_tokens == image_token_id).sum()
        video_nums = (vision_tokens == video_token_id).sum()
        input_tokens = input_ids.tolist()
        llm_pos_ids_list: list = []
        st = 0
        remain_images, remain_videos = image_nums, video_nums
        for _ in range(image_nums + video_nums):
            if image_token_id in input_tokens and remain_images > 0:
                ed_image = input_tokens.index(image_token_id, st)
            else:
                ed_image = len(input_tokens) + 1
            if video_token_id in input_tokens and remain_videos > 0:
                ed_video = input_tokens.index(video_token_id, st)
            else:
                ed_video = len(input_tokens) + 1
            if ed_image < ed_video:
                t, h, w = (
                    image_grid_thw[image_index][0],
                    image_grid_thw[image_index][1],
                    image_grid_thw[image_index][2],
                )
                second_per_grid_t = 0
                image_index += 1
                remain_images -= 1
                ed = ed_image
            else:
                t, h, w = (
                    video_grid_thw[video_index][0],
                    video_grid_thw[video_index][1],
                    video_grid_thw[video_index][2],
                )
                if second_per_grid_ts is not None:
                    second_per_grid_t = second_per_grid_ts[video_index]
                else:
                    second_per_grid_t = 1.0

                video_index += 1
                remain_videos -= 1
                ed = ed_video

            llm_grid_t, llm_grid_h, llm_grid_w = (
                t.item(),
                h.item() // spatial_merge_size,
                w.item() // spatial_merge_size,
            )
            text_len = ed - st

            st_idx = llm_pos_ids_list[-1].max() + 1 if len(llm_pos_ids_list) > 0 else 0
            llm_pos_ids_list.append(torch.arange(text_len).view(1, -1).expand(3, -1) + st_idx)

            t_index = torch.arange(llm_grid_t).view(-1, 1).expand(-1, llm_grid_h * llm_grid_w)
            t_index = (t_index * second_per_grid_t * tokens_per_second).long().flatten()
            h_index = torch.arange(llm_grid_h).view(1, -1, 1).expand(llm_grid_t, -1, llm_grid_w).flatten()
            w_index = torch.arange(llm_grid_w).view(1, 1, -1).expand(llm_grid_t, llm_grid_h, -1).flatten()
            llm_pos_ids_list.append(torch.stack([t_index, h_index, w_index]) + text_len + st_idx)
            st = ed + llm_grid_t * llm_grid_h * llm_grid_w

        if st < len(input_tokens):
            st_idx = llm_pos_ids_list[-1].max() + 1 if len(llm_pos_ids_list) > 0 else 0
            text_len = len(input_tokens) - st
            llm_pos_ids_list.append(torch.arange(text_len).view(1, -1).expand(3, -1) + st_idx)

        llm_positions = torch.cat(llm_pos_ids_list, dim=1).reshape(3, -1)
        position_ids[..., attention_mask == 1] = llm_positions.to(position_ids.device)
    else:
        if attention_mask is not None:
            position_ids = attention_mask.long().cumsum(-1) - 1
            position_ids.masked_fill_(attention_mask == 0, 1)
            position_ids = position_ids.unsqueeze(0).expand(3, -1).to(input_ids.device)
        else:
            position_ids = torch.arange(input_ids.shape[1], device=input_ids.device).view(1, -1).expand(3, -1)

    return position_ids