我有一個(gè)模型，我正在嘗試使用它。我正在解決這些錯(cuò)誤，但現(xiàn)在我認(rèn)為它已經(jīng)歸結(jié)為我層中的值。我收到此錯(cuò)誤：RuntimeError: Given groups=1, weight of size 24 1 3 3, expected input[512, 50, 50, 3] to have 1 channels, but got 50 channels instead我的參數(shù)是：LR = 5e-2N_EPOCHS = 30BATCH_SIZE = 512DROPOUT = 0.5我的圖像信息是：depth=24channels=3original height = 1600original width = 1200resized to 50x50這是我的數(shù)據(jù)的大?。篢rain shape (743, 50, 50, 3) (743, 7)Test shape (186, 50, 50, 3) (186, 7)Train pixels 0 255 188.12228712427097 61.49539262385051Test pixels 0 255 189.35559211469533 60.688278787628775我在這里試圖了解每一層的期望值，但是當(dāng)我在這里輸入它所說(shuō)的內(nèi)容時(shí)，https://towardsdatascience.com/pytorch-layer-dimensions-what-sizes-should-they-be-and-為什么-4265a41e01fd，它給了我關(guān)于錯(cuò)誤通道和內(nèi)核的錯(cuò)誤。我發(fā)現(xiàn) torch_summary 讓我對(duì)輸出有更多的了解，但它只會(huì)提出更多的問(wèn)題。這是我的 torch_summary 代碼：from torchvision import modelsfrom torchsummary import summaryimport torchimport torch.nn as nnclass CNN(nn.Module):    def __init__(self):        super(CNN, self).__init__()        self.conv1 = nn.Conv2d(1,24, kernel_size=5)  # output (n_examples, 16, 26, 26)        self.convnorm1 = nn.BatchNorm2d(24) # channels from prev layer        self.pool1 = nn.MaxPool2d((2, 2))  # output (n_examples, 16, 13, 13)        self.conv2 = nn.Conv2d(24,48,kernel_size=5)  # output (n_examples, 32, 11, 11)        self.convnorm2 = nn.BatchNorm2d(48) # 2*channels?        self.pool2 = nn.AvgPool2d((2, 2))  # output (n_examples, 32, 5, 5)        self.linear1 = nn.Linear(400,120)  # input will be flattened to (n_examples, 32 * 5 * 5)        self.linear1_bn = nn.BatchNorm1d(400) # features?        self.drop = nn.Dropout(DROPOUT)        self.linear2 = nn.Linear(400, 10)        self.act = torch.relu