注意:博主没有重写d2l的源代码文件,而是创建了一个新的python文件,并重写了该方法。
C:\Users\Administrator\anaconda3\envs\limu\python.exe G:/PyCharmProjects/limu-d2l/ch03/softmax_regression.py
Traceback (most recent call last):
File "" , line 1, in <module>
Traceback (most recent call last):
File "" , line 1, in <module>
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 116, in spawn_main
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 116, in spawn_main
exitcode = _main(fd, parent_sentinel)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 125, in _main
Traceback (most recent call last):
exitcode = _main(fd, parent_sentinel)
File "" , line 1, in <module>
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 125, in _main
Traceback (most recent call last):
File "" , line 1, in <module>
prepare(preparation_data)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 236, in prepare
prepare(preparation_data)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 236, in prepare
_fixup_main_from_path(data['init_main_from_path'])
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 287, in _fixup_main_from_path
_fixup_main_from_path(data['init_main_from_path'])
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 287, in _fixup_main_from_path
main_content = runpy.run_path(main_path,
File "C:\Users\Administrator\anaconda3\envs\limu\lib\runpy.py", line 265, in run_path
main_content = runpy.run_path(main_path,
File "C:\Users\Administrator\anaconda3\envs\limu\lib\runpy.py", line 265, in run_path
return _run_module_code(code, init_globals, run_name,
File "C:\Users\Administrator\anaconda3\envs\limu\lib\runpy.py", line 97, in _run_module_code
return _run_module_code(code, init_globals, run_name,
File "C:\Users\Administrator\anaconda3\envs\limu\lib\runpy.py", line 97, in _run_module_code
_run_code(code, mod_globals, init_globals,
File "C:\Users\Administrator\anaconda3\envs\limu\lib\runpy.py", line 87, in _run_code
_run_code(code, mod_globals, init_globals,
File "C:\Users\Administrator\anaconda3\envs\limu\lib\runpy.py", line 87, in _run_code
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 116, in spawn_main
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 116, in spawn_main
exec(code, run_globals)
File "G:\PyCharmProjects\limu-d2l\ch03\softmax_regression.py", line 97, in <module>
exec(code, run_globals)
File "G:\PyCharmProjects\limu-d2l\ch03\softmax_regression.py", line 97, in <module>
train_ch03(net, train_iter, test_iter, cross_entropy, num_epochs, updater)
File "G:\PyCharmProjects\limu-d2l\ch03\softmax_regression.py", line 81, in train_ch03
exitcode = _main(fd, parent_sentinel)
train_ch03(net, train_iter, test_iter, cross_entropy, num_epochs, updater)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 125, in _main
File "G:\PyCharmProjects\limu-d2l\ch03\softmax_regression.py", line 81, in train_ch03
exitcode = _main(fd, parent_sentinel)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 125, in _main
train_loss, train_acc = train_epoch_ch03(net, train_iter, loss, updater)train_loss, train_acc = train_epoch_ch03(net, train_iter, loss, updater)
File "G:\PyCharmProjects\limu-d2l\ch03\softmax_regression.py", line 64, in train_epoch_ch03
File "G:\PyCharmProjects\limu-d2l\ch03\softmax_regression.py", line 64, in train_epoch_ch03
prepare(preparation_data)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 236, in prepare
prepare(preparation_data)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 236, in prepare
for X, y in train_iter:for X, y in train_iter:
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 441, in __iter__
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 441, in __iter__
_fixup_main_from_path(data['init_main_from_path'])_fixup_main_from_path(data['init_main_from_path'])
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 287, in _fixup_main_from_path
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 287, in _fixup_main_from_path
main_content = runpy.run_path(main_path,main_content = runpy.run_path(main_path,
File "C:\Users\Administrator\anaconda3\envs\limu\lib\runpy.py", line 265, in run_path
File "C:\Users\Administrator\anaconda3\envs\limu\lib\runpy.py", line 265, in run_path
return self._get_iterator()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 388, in _get_iterator
return self._get_iterator()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 388, in _get_iterator
return _run_module_code(code, init_globals, run_name,return _run_module_code(code, init_globals, run_name,
File "C:\Users\Administrator\anaconda3\envs\limu\lib\runpy.py", line 97, in _run_module_code
File "C:\Users\Administrator\anaconda3\envs\limu\lib\runpy.py", line 97, in _run_module_code
return _MultiProcessingDataLoaderIter(self)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 1042, in __init__
return _MultiProcessingDataLoaderIter(self)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 1042, in __init__
_run_code(code, mod_globals, init_globals,
_run_code(code, mod_globals, init_globals,
File "C:\Users\Administrator\anaconda3\envs\limu\lib\runpy.py", line 87, in _run_code
File "C:\Users\Administrator\anaconda3\envs\limu\lib\runpy.py", line 87, in _run_code
exec(code, run_globals)
exec(code, run_globals)
File "G:\PyCharmProjects\limu-d2l\ch03\softmax_regression.py", line 97, in <module>
File "G:\PyCharmProjects\limu-d2l\ch03\softmax_regression.py", line 97, in <module>
train_ch03(net, train_iter, test_iter, cross_entropy, num_epochs, updater)
File "G:\PyCharmProjects\limu-d2l\ch03\softmax_regression.py", line 81, in train_ch03
train_ch03(net, train_iter, test_iter, cross_entropy, num_epochs, updater)
File "G:\PyCharmProjects\limu-d2l\ch03\softmax_regression.py", line 81, in train_ch03
w.start()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\process.py", line 121, in start
w.start()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\process.py", line 121, in start
train_loss, train_acc = train_epoch_ch03(net, train_iter, loss, updater)train_loss, train_acc = train_epoch_ch03(net, train_iter, loss, updater)
File "G:\PyCharmProjects\limu-d2l\ch03\softmax_regression.py", line 64, in train_epoch_ch03
File "G:\PyCharmProjects\limu-d2l\ch03\softmax_regression.py", line 64, in train_epoch_ch03
self._popen = self._Popen(self)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\context.py", line 224, in _Popen
for X, y in train_iter:for X, y in train_iter:
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 441, in __iter__
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 441, in __iter__
self._popen = self._Popen(self)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\context.py", line 224, in _Popen
return _default_context.get_context().Process._Popen(process_obj)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\context.py", line 327, in _Popen
return _default_context.get_context().Process._Popen(process_obj)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\context.py", line 327, in _Popen
return self._get_iterator()return self._get_iterator()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 388, in _get_iterator
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 388, in _get_iterator
return Popen(process_obj)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\popen_spawn_win32.py", line 45, in __init__
return Popen(process_obj) return _MultiProcessingDataLoaderIter(self)return _MultiProcessingDataLoaderIter(self)prep_data = spawn.get_preparation_data(process_obj._name)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 1042, in __init__
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 1042, in __init__
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 154, in get_preparation_data
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\popen_spawn_win32.py", line 45, in __init__
prep_data = spawn.get_preparation_data(process_obj._name)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 154, in get_preparation_data
_check_not_importing_main()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 134, in _check_not_importing_main
_check_not_importing_main()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 134, in _check_not_importing_main
raise RuntimeError('''
RuntimeError:
An attempt has been made to start a new process before the
current process has finished its bootstrapping phase.
This probably means that you are not using fork to start your
child processes and you have forgotten to use the proper idiom
in the main module:
if __name__ == '__main__':
freeze_support()
...
The "freeze_support()" line can be omitted if the program
is not going to be frozen to produce an executable.
w.start()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\process.py", line 121, in start
w.start()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\process.py", line 121, in start
raise RuntimeError('''
RuntimeError:
An attempt has been made to start a new process before the
current process has finished its bootstrapping phase.
This probably means that you are not using fork to start your
child processes and you have forgotten to use the proper idiom
in the main module:
if __name__ == '__main__':
freeze_support()
...
The "freeze_support()" line can be omitted if the program
is not going to be frozen to produce an executable.
self._popen = self._Popen(self)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\context.py", line 224, in _Popen
self._popen = self._Popen(self)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\context.py", line 224, in _Popen
return _default_context.get_context().Process._Popen(process_obj)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\context.py", line 327, in _Popen
return _default_context.get_context().Process._Popen(process_obj)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\context.py", line 327, in _Popen
return Popen(process_obj)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\popen_spawn_win32.py", line 45, in __init__
return Popen(process_obj)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\popen_spawn_win32.py", line 45, in __init__
prep_data = spawn.get_preparation_data(process_obj._name)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 154, in get_preparation_data
prep_data = spawn.get_preparation_data(process_obj._name)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 154, in get_preparation_data
_check_not_importing_main()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 134, in _check_not_importing_main
_check_not_importing_main()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\spawn.py", line 134, in _check_not_importing_main
raise RuntimeError('''
raise RuntimeError('''
RuntimeError: RuntimeError
An attempt has been made to start a new process before the
current process has finished its bootstrapping phase.
This probably means that you are not using fork to start your
child processes and you have forgotten to use the proper idiom
in the main module:
if __name__ == '__main__':
freeze_support()
...
The "freeze_support()" line can be omitted if the program
is not going to be frozen to produce an executable.:
An attempt has been made to start a new process before the
current process has finished its bootstrapping phase.
This probably means that you are not using fork to start your
child processes and you have forgotten to use the proper idiom
in the main module:
if __name__ == '__main__':
freeze_support()
...
The "freeze_support()" line can be omitted if the program
is not going to be frozen to produce an executable.
Traceback (most recent call last):
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 1132, in _try_get_data
data = self._data_queue.get(timeout=timeout)
File "C:\Users\Administrator\anaconda3\envs\limu\lib\multiprocessing\queues.py", line 108, in get
raise Empty
_queue.Empty
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "G:/PyCharmProjects/limu-d2l/ch03/softmax_regression.py", line 97, in <module>
train_ch03(net, train_iter, test_iter, cross_entropy, num_epochs, updater)
File "G:/PyCharmProjects/limu-d2l/ch03/softmax_regression.py", line 81, in train_ch03
train_loss, train_acc = train_epoch_ch03(net, train_iter, loss, updater)
File "G:/PyCharmProjects/limu-d2l/ch03/softmax_regression.py", line 64, in train_epoch_ch03
for X, y in train_iter:
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 633, in __next__
data = self._next_data()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 1328, in _next_data
idx, data = self._get_data()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 1294, in _get_data
success, data = self._try_get_data()
File "C:\Users\Administrator\anaconda3\envs\limu\lib\site-packages\torch\utils\data\dataloader.py", line 1145, in _try_get_data
raise RuntimeError('DataLoader worker (pid(s) {}) exited unexpectedly'.format(pids_str)) from e
RuntimeError: DataLoader worker (pid(s) 14032, 23312, 21048, 1952) exited unexpectedly
Process finished with exit code 1
这个错误是由于在使用多进程 DataLoader 时出现的问题,通常与 Windows 操作系统相关。在 Windows 上,使用多进程的 DataLoader 可能会导致一些问题,这与 Windows 的进程模型不太兼容。
在 Windows 上,将 DataLoader 的 num_workers
参数设置为 0,以使用单进程 DataLoader。这会禁用多进程加载数据,虽然可能会导致数据加载速度变慢,但通常可以解决与多进程 DataLoader 相关的问题。
def get_dataloader_workers():
"""Use 4 processes to read the data.
Defined in :numref:`sec_utils`"""
return 4
def load_data_fashion_mnist(batch_size, resize=None):
"""Download the Fashion-MNIST dataset and then load it into memory.
Defined in :numref:`sec_utils`"""
trans = [transforms.ToTensor()]
if resize:
trans.insert(0, transforms.Resize(resize))
trans = transforms.Compose(trans)
mnist_train = torchvision.datasets.FashionMNIST(
root="../data", train=True, transform=trans, download=True)
mnist_test = torchvision.datasets.FashionMNIST(
root="../data", train=False, transform=trans, download=True)
return (torch.utils.data.DataLoader(mnist_train, batch_size, shuffle=True,
num_workers=get_dataloader_workers()),
torch.utils.data.DataLoader(mnist_test, batch_size, shuffle=False,
num_workers=get_dataloader_workers()))
def load_data_fashion_mnist(batch_size, resize=None, num_workers=4):
"""下载Fashion-MNIST数据集,然后将其加载到内存中"""
trans = [transforms.ToTensor()]
if resize:
trans.index(0, transforms.Resize(resize))
trans = transforms.Compose(trans)
mnist_train = torchvision.datasets.FashionMNIST(root='../data', train=True, transform=trans, download=True)
mnist_test = torchvision.datasets.FashionMNIST(root='../data', train=False, transform=trans, download=True)
return (data.DataLoader(mnist_train, batch_size, shuffle=True, num_workers=num_workers),
data.DataLoader(mnist_test, batch_size, shuffle=False, num_workers=num_workers))
train_iter, test_iter = fashion_mnist.load_data_fashion_mnist(batch_size, num_workers=0)
在 PyTorch 的 DataLoader 中,num_workers
参数控制了用于加载数据的子进程数量。当 num_workers
被设置为 0 时,实际上是表示不使用任何子进程来加载数据,即单进程加载数据。
为什么不是零进程?这是因为 DataLoader 需要至少一个进程来加载数据,这个进程被称为主进程。主进程负责数据加载和分发给训练的进程。当 num_workers
设置为 0 时,只有主进程用于加载和处理数据,没有额外的子进程。这是一种单进程的数据加载方式。
如果将 num_workers
设置为 1,则会有一个额外的子进程来加载数据,总共会有两个进程:一个主进程和一个数据加载子进程。这种设置可以在某些情况下提高数据加载的效率,特别是当数据加载耗时较长时,子进程可以并行地加载数据,从而加速训练过程。
import torch
from d2l import torch as d2l
import fashion_mnist
batch_size = 256
train_iter, test_iter = fashion_mnist.load_data_fashion_mnist(batch_size, num_workers=0)
# 初始化模型参数
num_inputs = 784 # 每个输入图像的通道数为1, 高度和宽度均为28像素
num_outputs = 10
W = torch.normal(0, 0.01, size=(num_inputs, num_outputs), requires_grad=True)
b = torch.zeros(num_outputs, requires_grad=True)
# 定义softmax操作
def softmax(X):
"""
矩阵中的非常大或非常小的元素可能造成数值上溢或者下溢
解决方案: P84 3.7.2 重新审视softmax的实现
"""
X_exp = torch.exp(X)
partition = X_exp.sum(1, keepdim=True)
return X_exp / partition
# 定义模型
def net(X):
return softmax(torch.matmul(X.reshape((-1, W.shape[0])), W) + b)
# 定义损失函数
def cross_entropy(y_hat, y):
return - torch.log(y_hat[range(len(y_hat)), y])
# 分类精度
def accuracy(y_hat, y):
"""计算预测正确的数量"""
if len(y_hat.shape) > 1 and y_hat.shape[1] > 1:
y_hat = y_hat.argmax(axis=1)
cmp = y_hat.type(y.dtype) == y
return float(cmp.type(y.dtype).sum())
def evaluate_accuracy(net, data_iter):
"""计算在制定数据集上模型的精度"""
if isinstance(net, torch.nn.Module):
net.eval()
metric = d2l.Accumulator(2)
with torch.no_grad():
for X, y in data_iter:
metric.add(accuracy(net(X), y), y.numel())
return metric[0] / metric[1]
# 训练
def train_epoch_ch03(net, train_iter, loss, updater):
if isinstance(net, torch.nn.Module):
net.train()
# 训练损失总和, 训练准确度总和, 样本数
metric = d2l.Accumulator(3)
for X, y in train_iter:
y_hat = net(X)
l = loss(y_hat, y)
if isinstance(updater, torch.optim.Optimizer):
updater.zero_grad()
l.mean().backward()
updater.step()
else:
l.sum().backward()
updater(X.shape[0])
metric.add(float(l.sum()), accuracy(y_hat, y), y.numel())
# 返回训练损失和训练精度
return metric[0] / metric[2], metric[1] / metric[2]
def train_ch03(net, train_iter, test_iter, loss, num_epochs, updater):
for epoch in range(num_epochs):
train_loss, train_acc = train_epoch_ch03(net, train_iter, loss, updater)
test_acc = evaluate_accuracy(net, test_iter)
print(f'epoch {epoch + 1}, train_loss {train_loss:f}, train_acc {train_acc:f}, test_acc {test_acc:f}')
assert train_loss < 0.5, train_loss
assert train_acc <= 1 and train_acc > 0.7, train_acc
assert test_acc <= 1 and test_acc > 0.7, test_acc
lr = 0.1
def updater(batch_size):
return d2l.sgd([W, b], lr, batch_size)
num_epochs = 10
train_ch03(net, train_iter, test_iter, cross_entropy, num_epochs, updater)