delta-mpc / delta-node

Delta node receives Delta tasks, distributes them across the network and executes tasks from the network.
https://deltampc.com
Apache License 2.0
4 stars 2 forks source link

[bug] get 413 http error status when use big model #35

Open TYTaO opened 2 years ago

TYTaO commented 2 years ago

Describe the bug 当使用较大的模型训练数据时,创建任务会报错 3423047cdef424209110149ee5de46a

Expected behavior 使用模型如下:

class CNN_FEMNIST(nn.Module):
    def __init__(self, only_digits=False):
        super(CNN_FEMNIST, self).__init__()
        self.conv2d_1 = nn.Conv2d(1, 32, kernel_size=3)
        self.max_pooling = nn.MaxPool2d(2, stride=2)
        self.conv2d_2 = nn.Conv2d(32, 64, kernel_size=3)
        self.dropout_1 = nn.Dropout(0.25)
        self.flatten = nn.Flatten()
        self.linear_1 = nn.Linear(9216, 128)
        self.dropout_2 = nn.Dropout(0.5)
        self.linear_2 = nn.Linear(128, 10 if only_digits else 62)
        self.relu = nn.ReLU()
        # self.softmax = nn.Softmax(dim=1)

    def forward(self, x):
        x = self.conv2d_1(x)
        x = self.relu(x)
        x = self.conv2d_2(x)
        x = self.relu(x)
        x = self.max_pooling(x)
        x = self.dropout_1(x)
        x = self.flatten(x)
        x = self.linear_1(x)
        x = self.relu(x)
        x = self.dropout_2(x)
        x = self.linear_2(x)
        # x = self.softmax(x)
        return x
TYTaO commented 2 years ago

通过修改dashboard里nginx里的配置可以解决,https://serverfault.com/questions/814767/413-request-entity-too-large-in-nginx-with-client-max-body-size-set

mh739025250 commented 2 years ago

感谢回复。 我觉得修改nginx配置不能从根本上解决这个问题。因为实际部署时,可以不通过dashboard中的nginx,直接调用delta-node的api,这时候网关的配置是不受控制的。

TYTaO commented 2 years ago

感谢回复。 我觉得修改nginx配置不能从根本上解决这个问题。因为实际部署时,可以不通过dashboard中的nginx,直接调用delta-node的api,这时候网关的配置是不受控制的。

嗯嗯,测试时不用nginx是不会收到这个限制的