接下来安装pytorch 

Previous PyTorch Versions

---

视频抽帧 

import cv2def extract_frames(video_path, output_folder, frame_rate=1):"""从视频中抽取帧。:param video_path: 视频文件的路径:param output_folder: 存储帧的文件夹路径:param frame_rate: 抽取的帧率，默认为1（每秒1帧）"""# 打开视频文件cap = cv2.VideoCapture(video_path)if not cap.isOpened():print("Error: Could not open video.")return# 获取视频的帧率，这对于控制抽取帧的间隔很有用fps = cap.get(cv2.CAP_PROP_FPS)frame_interval = int(fps / frame_rate)  # 计算间隔帧数frame_count = 0frame_index = 0# 循环读取帧直到视频结束while True:ret, frame = cap.read()if not ret:break  # 如果读取失败，则退出循环# 根据帧率间隔抽取帧if frame_count % frame_interval == 0:frame_filename = f"{output_folder}/frame_{frame_index:05d}.jpg"cv2.imwrite(frame_filename, frame)  # 保存帧print(f"Saved {frame_filename}")frame_index += 1frame_count += 1# 释放视频捕获对象cap.release()print("Extraction complete.")# 使用示例
video_path = '20035.MP4'  # 视频文件路径
output_folder = 'imgs'  # 输出文件夹路径
extract_frames(video_path, output_folder, frame_rate=2)  # 每秒抽取2帧

---