国产片侵犯亲女视频播放_亚洲精品二区_在线免费国产视频_欧美精品一区二区三区在线_少妇久久久_在线观看av不卡

腳本之家,腳本語言編程技術及教程分享平臺!
分類導航

Python|VBS|Ruby|Lua|perl|VBA|Golang|PowerShell|Erlang|autoit|Dos|bat|

服務器之家 - 腳本之家 - Python - PyTorch中的拷貝與就地操作詳解

PyTorch中的拷貝與就地操作詳解

2021-08-11 00:36Chris_34 Python

這篇文章主要給大家介紹了關于PyTorch中拷貝與就地操作的相關資料,文中通過示例代碼介紹的非常詳細,對大家的學習或者工作具有一定的參考學習價值,需要的朋友們下面隨著小編來一起學習學習吧

前言

PyTroch中我們經常使用到Numpy進行數據的處理,然后再轉為Tensor,但是關系到數據的更改時我們要注意方法是否是共享地址,這關系到整個網絡的更新。本篇就In-palce操作,拷貝操作中的注意點進行總結。

In-place操作

pytorch中原地操作的后綴為_,如.add_()或.scatter_(),就地操作是直接更改給定Tensor的內容而不進行復制的操作,即不會為變量分配新的內存。Python操作類似+=或*=也是就地操作。(我加了我自己~)

為什么in-place操作可以在處理高維數據時可以幫助減少內存使用呢,下面使用一個例子進行說明,定義以下簡單函數來測量PyTorch的異位ReLU(out-of-place)和就地ReLU(in-place)分配的內存:

?
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
import torch # import main library
import torch.nn as nn # import modules like nn.ReLU()
import torch.nn.functional as F # import torch functions like F.relu() and F.relu_()
 
def get_memory_allocated(device, inplace = False):
 '''
 Function measures allocated memory before and after the ReLU function call.
 INPUT:
 - device: gpu device to run the operation
 - inplace: True - to run ReLU in-place, False - for normal ReLU call
 '''
 
 # Create a large tensor
 t = torch.randn(10000, 10000, device=device)
 
 # Measure allocated memory
 torch.cuda.synchronize()
 start_max_memory = torch.cuda.max_memory_allocated() / 1024**2
 start_memory = torch.cuda.memory_allocated() / 1024**2
 
 # Call in-place or normal ReLU
 if inplace:
 F.relu_(t)
 else:
 output = F.relu(t)
 
 # Measure allocated memory after the call
 torch.cuda.synchronize()
 end_max_memory = torch.cuda.max_memory_allocated() / 1024**2
 end_memory = torch.cuda.memory_allocated() / 1024**2
 
 # Return amount of memory allocated for ReLU call
 return end_memory - start_memory, end_max_memory - start_max_memory
 # setup the device
device = torch.device('cuda:0' if torch.cuda.is_available() else "cpu")
#開始測試
# Call the function to measure the allocated memory for the out-of-place ReLU
memory_allocated, max_memory_allocated = get_memory_allocated(device, inplace = False)
print('Allocated memory: {}'.format(memory_allocated))
print('Allocated max memory: {}'.format(max_memory_allocated))
'''
Allocated memory: 382.0
Allocated max memory: 382.0
'''
#Then call the in-place ReLU as follows:
memory_allocated_inplace, max_memory_allocated_inplace = get_memory_allocated(device, inplace = True)
print('Allocated memory: {}'.format(memory_allocated_inplace))
print('Allocated max memory: {}'.format(max_memory_allocated_inplace))
'''
Allocated memory: 0.0
Allocated max memory: 0.0
'''

看起來,使用就地操作可以幫助我們節省一些GPU內存。但是,在使用就地操作時應該格外謹慎。

就地操作的主要缺點主要原因有2點,官方文檔

1.可能會覆蓋計算梯度所需的值,這意味著破壞了模型的訓練過程。

2.每個就地操作實際上都需要實現來重寫計算圖。異地操作Out-of-place分配新對象并保留對舊圖的引用,而就地操作則需要更改表示此操作的函數的所有輸入的創建者。

在Autograd中支持就地操作很困難,并且在大多數情況下不鼓勵使用。Autograd積極的緩沖區釋放和重用使其非常高效,就地操作實際上降低內存使用量的情況很少。除非在沉重的內存壓力下運行,否則可能永遠不需要使用它們。

總結:Autograd很香了,就地操作要慎用。

拷貝方法

淺拷貝方法: 共享 data 的內存地址,數據會同步變化

* a.numpy() # Tensor—>Numpy array

* view() #改變tensor的形狀,但共享數據內存,不要直接使用id進行判斷

* y = x[:] # 索引

* torch.from_numpy() # Numpy array—>Tensor

* torch.detach() # 新的tensor會脫離計算圖,不會牽扯梯度計算。

* model:forward()

還有很多選擇函數也是數據共享內存,如index_select() masked_select() gather()。

以及后文提到的就地操作in-place。

深拷貝方法:

* torch.clone() # 新的tensor會保留在計算圖中,參與梯度計算

下面進行驗證,首先驗證淺拷貝:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
import torch as t
import numpy as np
a = np.ones(4)
b = t.from_numpy(a) # Numpy->Tensor
print(a)
print(b)
'''輸出:
[1. 1. 1. 1.]
tensor([1., 1., 1., 1.], dtype=torch.float64)
'''
b.add_(1)# add_會修改b自身
print(a)
print(b)
'''輸出:
[2. 2. 2. 2.]
tensor([2., 2., 2., 2.], dtype=torch.float64)
b進行add操作后, a,b同步發生了變化
'''

Tensor和numpy對象共享內存(淺拷貝操作),所以他們之間的轉換很快,且會同步變化。

造torch中y = x + y這樣的運算是會新開內存的,然后將y指向新內存。為了進行驗證,我們可以使用Python自帶的id函數:如果兩個實例的ID一致,那么它們所對應的內存地址相同;但需要注意是在torch中還有些特殊,數據共享時直接打印tensor的id仍然會出現不同。

?
1
2
3
4
5
6
x = torch.tensor([1, 2])
y = torch.tensor([3, 4])
id_0 = id(y)
y = y + x
print(id(y) == id_0)
# False

這時使用索引操作不會開辟新的內存,而想指定結果到原來的y的內存,我們可以使用索引來進行替換操作。比如把x + y的結果通過[:]寫進y對應的內存中。

?
1
2
3
4
5
6
x = torch.tensor([1, 2])
y = torch.tensor([3, 4])
id_0 = id(y)
y[:] = y + x
print(id(y) == id_0)
# True

另外,以下兩種方式也可以索引到相同的內存:

  • torch.add(x, y, out=y)
  • y += x, y.add_(x)
?
1
2
3
4
5
6
7
x = torch.tensor([1, 2])
y = torch.tensor([3, 4])
id_0 = id(y)
torch.add(x, y, out=y)
# y += x, y.add_(x)
print(id(y) == id_0)
# True

clone() 與 detach() 對比

Torch 為了提高速度,向量或是矩陣的賦值是指向同一內存的,這不同于 Matlab。如果需要保存舊的tensor即需要開辟新的存儲地址而不是引用,可以用 clone() 進行深拷貝,

首先我們來打印出來clone()操作后的數據類型定義變化:

(1). 簡單打印類型

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
import torch
 
a = torch.tensor(1.0, requires_grad=True)
b = a.clone()
c = a.detach()
a.data *= 3
b += 1
 
print(a) # tensor(3., requires_grad=True)
print(b)
print(c)
 
'''
輸出結果:
tensor(3., requires_grad=True)
tensor(2., grad_fn=<AddBackward0>)
tensor(3.)  # detach()后的值隨著a的變化出現變化
'''

grad_fn=<CloneBackward>,表示clone后的返回值是個中間變量,因此支持梯度的回溯。clone操作在一定程度上可以視為是一個identity-mapping函數。

detach()操作后的tensor與原始tensor共享數據內存,當原始tensor在計算圖中數值發生反向傳播等更新之后,detach()的tensor值也發生了改變。

注意: 在pytorch中我們不要直接使用id是否相等來判斷tensor是否共享內存,這只是充分條件,因為也許底層共享數據內存,但是仍然是新的tensor,比如detach(),如果我們直接打印id會出現以下情況。

?
1
2
3
4
5
6
7
8
import torch as t
a = t.tensor([1.0,2.0], requires_grad=True)
b = a.detach()
#c[:] = a.detach()
print(id(a))
print(id(b))
#140568935450520
140570337203616

顯然直接打印出來的id不等,我們可以通過簡單的賦值后觀察數據變化進行判斷。

(2). clone()的梯度回傳

detach()函數可以返回一個完全相同的tensor,與舊的tensor共享內存,脫離計算圖,不會牽扯梯度計算。

而clone充當中間變量,會將梯度傳給源張量進行疊加,但是本身不保存其grad,即值為None

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
import torch
a = torch.tensor(1.0, requires_grad=True)
a_ = a.clone()
y = a**2
z = a ** 2+a_ * 3
y.backward()
print(a.grad) # 2
z.backward()
print(a_.grad)   # None. 中間variable,無grad
print(a.grad)
'''
輸出:
tensor(2.)
None
tensor(7.) # 2*2+3=7
'''

使用torch.clone()獲得的新tensor和原來的數據不再共享內存,但仍保留在計算圖中,clone操作在不共享數據內存的同時支持梯度梯度傳遞與疊加,所以常用在神經網絡中某個單元需要重復使用的場景下。

通常如果原tensor的requires_grad=True,則:

  • clone()操作后的tensor requires_grad=True
  • detach()操作后的tensor requires_grad=False。
?
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
import torch
torch.manual_seed(0)
 
x= torch.tensor([1., 2.], requires_grad=True)
clone_x = x.clone()
detach_x = x.detach()
clone_detach_x = x.clone().detach()
 
f = torch.nn.Linear(2, 1)
y = f(x)
y.backward()
 
print(x.grad)
print(clone_x.requires_grad)
print(clone_x.grad)
print(detach_x.requires_grad)
print(clone_detach_x.requires_grad)
'''
輸出結果如下:
tensor([-0.0053, 0.3793])
True
None
False
False
'''

另一個比較特殊的是當源張量的 require_grad=False,clone后的張量 require_grad=True,此時不存在張量回傳現象,可以得到clone后的張量求導。

如下:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
import torch
a = torch.tensor(1.0)
a_ = a.clone()
a_.requires_grad_() #require_grad=True
y = a_ ** 2
y.backward()
print(a.grad) # None
print(a_.grad)
'''
輸出:
None
tensor(2.)
'''

總結:

torch.detach() —新的tensor會脫離計算圖,不會牽扯梯度計算

torch.clone() — 新的tensor充當中間變量,會保留在計算圖中,參與梯度計算(回傳疊加),但是一般不會保留自身梯度。

原地操作(in-place, such as resize_ / resize_as_ / set_ / transpose_) 在上面兩者中執行都會引發錯誤或者警告。

引用官方文檔的話:如果你使用了in-place operation而沒有報錯的話,那么你可以確定你的梯度計算是正確的。另外盡量避免in-place的使用。

到此這篇關于PyTorch中拷貝與就地操作的文章就介紹到這了,更多相關PyTorch拷貝與就地操作內容請搜索服務器之家以前的文章或繼續瀏覽下面的相關文章希望大家以后多多支持服務器之家!

原文鏈接:https://blog.csdn.net/weixin_43199584/article/details/106770647

延伸 · 閱讀

精彩推薦
主站蜘蛛池模板: 国产成人一区 | 亚洲欧洲在线观看 | 中文字幕精品一区二区三区精品 | 曰韩免费视频 | 天天操综合网 | 精品中文在线 | 超级黄色毛片 | 国产成人久久av免费高清密臂 | 午夜影院在线 | 精品国产精品三级精品av网址 | 中文字幕在线观看一区 | 亚洲精品1区 | 久久av一区二区三区 | 色网网站 | 成人a级片在线观看 | 久久久国产一区 | 国产综合免费视频 | 91仓库 | 国产福利在线观看 | 久久精品视 | 黄色片免费在线看 | 亚洲骚片 | 亚洲精品免费在线 | 丝袜+亚洲+另类+欧美+变态 | 亚洲国产精品激情在线观看 | 国产日韩欧美高清 | 激情毛片 | 天天天干天天天操 | 99精品视频免费观看 | 久久伊99综合婷婷久久伊 | 欧美国产一区二区 | 色成人亚洲www78ixcom | 日韩在线观看中文字幕 | 欧美黑人性暴力猛交喷水黑人巨大 | 日韩精品www | 亚洲电影在线 | 欧美激情在线观看 | 亚洲精彩视频在线 | 中文字幕乱码一区二区三区 | 精品护士一区二区三区 | 亚洲免费在线视频 |