tensorboard 出现的问题

首先要安装

pip install tensorboard
 

运行命令为 tensorboard --logdir=日志文件的绝对地址  --port=8008  端口号随意

如果运行上述命令出现以下问题

TensorFlow installation not found - running with reduced feature set.

说明TensorFlow 没有安装

pip install tensorflow
 

可视化模型结构

model = SwinUNETR(
    img_size=(args.roi_x, args.roi_y, args.roi_z),
    in_channels=args.in_channels,
    out_channels=args.out_channels,
    feature_size=args.feature_size,
    use_checkpoint=args.use_checkpoint,
)
# print(model)
input = torch.rand((1, 4, 128, 128, 128))   这里是模型的输入
writer = SummaryWriter("保存的路径")
writer.add_graph(model, input_to_model=torch.rand(1, 4, 128, 128, 128))
writer.close()

上面定义模型，

tensorboard --logdir=保存的路径 --port=8008