查看网络结构
饿了就干饭
经常更新NLP、深度学习相关内容,感兴趣的话可以关注哈~
展开
-
nn.Sequential(*modules)
将写好的模型进行修剪再结合创建模型model = 巴拉巴拉modules = list(model.children())# 可以添加索引[] 用于切片swintransformer = nn.Sequential(*modules)print(swintransformer)Sequential( (0): PatchEmbed( (proj): Conv2d(3, 96, kernel_size=(4, 4), stride=(4, 4)) (norm): LayerNo原创 2022-01-06 16:27:03 · 2055 阅读 · 0 评论 -
记一次灵活的模型训练生成的pth转onnx文件失败
代码import torchimport torchvisiondevice = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")dummy_input = torch.randn(1, 3, 224, 224).to(device)model = torch.load(r'D:\jiaoyidi\swin_transformer_flower\weights\model---0.pth')model.eval()in原创 2022-01-05 22:55:48 · 1769 阅读 · 3 评论 -
查看swintransformer的网络结构.children()
import osimport jsonimport torchfrom PIL import Imagefrom torchvision import transformsimport matplotlib.pyplot as plt#导入对应的模型from model import swin_tiny_patch4_window7_224 as create_modeldevice = torch.device("cuda:0" if torch.cuda.is_available()原创 2022-01-05 11:44:38 · 228 阅读 · 2 评论 -
查看resnet101的网络结构
代码import torchfrom torch import nnimport torchvisiondevice = torch.device("cuda" if torch.cuda.is_available() else "cpu")resnet = torchvision.models.resnet101(pretrained=True) # pretrained ImageNet ResNet-101print(resnet)输出:ResNet( (conv1): Con原创 2022-01-04 16:37:38 · 3879 阅读 · 0 评论