File size: 6,335 Bytes
bbaf249
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
from time_r1.utils.qwen_vl_utils_prepare import floor_by_factor, FRAME_FACTOR, smart_resize, ceil_by_factor
import decord
import torch
import os
import tqdm
import glob
import multiprocessing
from torchvision import io, transforms
from torchvision.transforms import InterpolationMode
from functools import partial
from time_r1.utils.io import load_jsonl
import debugpy
# try:
#     # 5678 is the default attach port in the VS Code debug configurations. Unless a host and port are specified, host defaults to 127.0.0.1
#     debugpy.listen(("localhost", 9511))
#     print("Waiting for debugger attach")
#     debugpy.wait_for_client()
# except Exception as e:
#     pass



FPS_MIN_FRAMES = 4
FPS_MAX_FRAMES = 120  # 建议根据你的显存情况调整,60-80 是常用范围
FRAME_FACTOR = 2
def get_video_tensor(video_path, target_fps=2, image_factor = 28, min_pixels = 28 * 28 * 4, max_pixels = 28 * 28 * 128):
    """
    将视频以固定帧率提前抽帧、解码保存为tensor,用于后续训练
    """
    vr = decord.VideoReader(video_path)
    total_frames, video_fps = len(vr), vr.get_avg_fps()


    min_nframes = ceil_by_factor(FPS_MIN_FRAMES, FRAME_FACTOR) # :ceil_by_factor 是为了保证帧数是 FRAME_FACTOR 的倍数
    max_frames = floor_by_factor(min(FPS_MAX_FRAMES, total_frames), FRAME_FACTOR)
    nframes = total_frames / video_fps * target_fps # 理论帧数
    if nframes > total_frames:
        print(f"smart_nframes: nframes[{nframes}] > total_frames[{total_frames}]")

    nframes = min(min(max(nframes, min_nframes), max_frames), total_frames)
    nframes = floor_by_factor(nframes, FRAME_FACTOR)
 
    frame_idx = torch.linspace(0, total_frames - 1, nframes).round().long().tolist() # 均匀采样
    frame_tensor = vr.get_batch(frame_idx).asnumpy()
    frame_tensor = torch.tensor(frame_tensor).permute(0, 3, 1, 2)  # Convert to TCHW format
    sample_fps = nframes / max(total_frames, 1e-6) * video_fps
    height, width = frame_tensor.shape[2], frame_tensor.shape[3]
    resized_height, resized_width = smart_resize(
        height,
        width,
        factor=image_factor,
        min_pixels=min_pixels,
        max_pixels=max_pixels,
    )
    frame_tensor = transforms.functional.resize(
        frame_tensor,
        [resized_height, resized_width],
        interpolation=InterpolationMode.BICUBIC,
        antialias=True,
    )
    frame_cache = {
        "frame_tensor": frame_tensor,
        "fps": sample_fps,
    }
    return frame_cache


def process_single_video(video_path, target_fps=1, image_factor = 28, min_pixels = 28 * 28 * 128, max_pixels = 28 * 28 * 256):
    save_dir = '/data/shuimu.chen/LongVideoBench/video_cache_300p_120'
    os.makedirs(save_dir, exist_ok=True)
    video_filename = os.path.basename(video_path)
    save_path = os.path.join(save_dir, video_filename + '.frame_cache')
    """Helper function to process and save frame cache for a single video."""
    print(f"Processing {video_path}...")
    try:
        frame_cache = get_video_tensor(video_path, target_fps, image_factor, min_pixels, max_pixels)
        torch.save(frame_cache, save_path)
        print(f"Successfully saved frame cache for {video_path}")
    except Exception as e:
        print(f"Error processing {video_path}: {e}")


# def process_single_video(video_path, target_fps=1, image_factor = 28, min_pixels = 28 * 28 * 128, max_pixels = 28 * 28 * 256):
#     save_path = '/data/shuimu.chen/LongVideoBench/video_cache_3p_60'
#     """Helper function to process and save frame cache for a single video."""
#     print(f"Processing {video_path}...")
#     try:
#         frame_cache = get_video_tensor(video_path, target_fps, image_factor, min_pixels, max_pixels)
#         torch.save(frame_cache, video_path + ".frame_cache")
#         print(f"Successfully saved frame cache for {video_path}")
#     except Exception as e:
#         print(f"Error processing {video_path}: {e}")


def prepare_frame_cache(video_root, dataset_path=None, num_workers=8, target_fps=1, overwrite=False, image_factor = 28, min_pixels = 28 * 28 * 128, max_pixels = 28 * 28 * 256):
    if dataset_path is not None:
        # 修改这里:直接使用json.load而不是load_jsonl
        import json
        with open(dataset_path, 'r', encoding='utf-8') as f:
            video_data = json.load(f)  # 这是JSON数组
        
        # 提取video_path字段,并去重(同一个视频可能有多条记录)
        video_paths = set()  # 使用set去重
        for v in video_data:
            if "video_path" in v:
                video_paths.add(v["video_path"])
            elif "video" in v:
                # 如果有video字段,可能需要拼接路径
                video_path = os.path.join(video_root, v["video"])
                video_paths.add(video_path)
        
        video_list = list(video_paths)
    else:
        video_list = glob.glob(os.path.join(video_root, "*.mp4"))
    if not video_list:
        print(f"No MP4 videos found in {video_root}")
        return
    # remove videos that already have frame cache
    if not overwrite:
        print("skipping videos that already have frame cache")
        num_total = len(video_list)
        video_list = [v for v in video_list if not os.path.exists(v + ".frame_cache")]
        num_skipped = num_total - len(video_list)
        print(f"skipped {num_skipped} videos")

    if num_workers is None:
        num_workers = multiprocessing.cpu_count()  # Default to using all available CPU cores
    
    print(f"Found {len(video_list)} videos. Starting processing with {num_workers} workers...")

    # Use a multiprocessing Pool to process videos in parallel
    with multiprocessing.Pool(processes=num_workers) as pool:
        # Using tqdm with pool.imap_unordered for progress bar and efficient iteration
        # We wrap process_single_video if it needs more arguments or if we want to handle results
        # For this case, process_single_video only takes video_path
        func = partial(process_single_video, target_fps=target_fps, image_factor = image_factor, min_pixels = min_pixels, max_pixels = max_pixels)
        list(tqdm.tqdm(pool.imap_unordered(func, video_list), total=len(video_list)))

    print("All videos processed.")


if __name__ == "__main__":
    import fire
    fire.Fire(prepare_frame_cache)