python 多进程multiprocessing 内存占用 物理内存 虚拟内存 windows 分析

 

只有

import time
time.sleep(10)

 

前边加上

from torch.utils.data import Dataset

 

 

 

import torch# 检查是否支持 CUDA
if torch.cuda.is_available():print("CUDA is available!")print(f"Device count: {torch.cuda.device_count()}")print(f"Current device: {torch.cuda.current_device()}")print(f"Device name: {torch.cuda.get_device_name(torch.cuda.current_device())}")
else:print("CUDA is not available.")import time
time.sleep(10)

 

这样看来

import torch和from torch.utils.data import Dataset占用的内存是一样的

 

 

 

from transformers import AutoModelForZeroShotImageClassification, AutoProcessor# import torch
#
# # 检查是否支持 CUDA
# if torch.cuda.is_available():
#     print("CUDA is available!")
#     print(f"Device count: {torch.cuda.device_count()}")
#     print(f"Current device: {torch.cuda.current_device()}")
#     print(f"Device name: {torch.cuda.get_device_name(torch.cuda.current_device())}")
# else:
#     print("CUDA is not available.")
import time
time.sleep(10)

中间的torch取消注释后几乎不变

 

 

 

多进程测试

from transformers import AutoModelForZeroShotImageClassification, AutoProcessor
import timedef apply_test(s):time.sleep(3)print(s)if __name__ == '__main__':import multiprocessingpool = multiprocessing.Pool(3)for i in range(3):res = pool.apply_async(apply_test, [i])time.sleep(20)

 

 

 

 

import timedef apply_test(s):time.sleep(3)print(s)if __name__ == '__main__':from transformers import AutoModelForZeroShotImageClassification, AutoProcessorimport multiprocessingpool = multiprocessing.Pool(3)for i in range(3):res = pool.apply_async(apply_test, [i])time.sleep(20)