您好，我想问一下怎么使用single_test.py推理视频

[wangshiwei567]

您好，我想问一下怎么使用single_test.py推理视频

[Fafa-DL]

这个问题以前有人私信问过我，当时手把手教他改了，你只需要读取视频为图片序列后再逐一检测和显示就行，如果需要的话我还是上传一个视频检测文件吧

[wangshiwei567]

我已经读取了视频，并传给了inference_model，但出现了报错，找不到解决方法。
`import time
from argparse import ArgumentParser
import os
import sys
import cv2
sys.path.insert(0, os.getcwd())
import torch

from utils.inference import inference_model, init_model, show_result_pyplot
from utils.train_utils import get_info, file2dict
from models.build import BuildNet

def main():

video = '/home/sw/PycharmProjects/data/2021115163527.mp4'
config = '/home/sw/PycharmProjects/Classification/Awesome-Backbones-0.6.0/models/shufflenet/shufflenet_v2.py'
device = 'cuda:0'
save_path = './log'
classes_map = '../datas/annotations.txt'

classes_names, _ = get_info(classes_map)
# build the model from a config file and a checkpoint file
model_cfg, train_pipeline, val_pipeline, data_cfg, lr_config, optimizer_cfg = file2dict(config)
if device is not None:
    device = torch.device(device)
else:
    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

model = BuildNet(model_cfg)
model = init_model(model, data_cfg, device=device, mode='eval')

cap=cv2.VideoCapture(video)
while True:
    flag, frame = cap.read()
    if not flag:
        break
    t1 = time.time()
    # test a single image
    frame = torch.from_numpy(frame)
    result = inference_model(model, frame, val_pipeline, classes_names)

    show_result_pyplot(model, img, result, out_file=save_path)

if name == 'main':
main()`

[Fafa-DL]

已更新，请拉取最新版本，或者单独下载tools/video_test.py以及更新utils/inference.py，如果解决了你的问题还麻烦在这里给个回复哈

[wangshiwei567]

您好，已经可以推理视频了，非常感谢。