Google Colab上的Tensorboard返回链接，但不执行任何操作

Question

我一直在使用Google Colab进行Tensorflow API进行对象检测，目前，该模型已经过训练。我需要查看Tensorboard进行培训。

这是培训电话：

!python 'Object Detection/models/research/object_detection/legacy/train.py' --logtostderr --train_dir='Object Detection/Model'/ --pipeline_config_path='Object Detection/Faster_RCNN.config'

我尝试了以下命令，这些是它们的结果：

1. !tensorboard --logdir='Object Detection/Model'

   TensorBoard 1.15.0 at http://cca84e64bb22:6006/ (Press CTRL+C to quit)

2. %tensorboard --logdir='Object Detection/Model'

   显示Launching Tensorboard...几秒钟并结束

3. %load_ext tensorboard

   The tensorboard extension is already loaded. To reload it, use:
     %reload_ext tensorboard

4. %reload_ext tensorboard

   空白-无输出

5. 我从'tensorboard_in_notebooks.ipynb'中找到了这个

from tensorboard import notebook
notebook.list()
notebook.display(port=6006, height=1000) 

Known TensorBoard instances:
    - port 6006: logdir /Object Detection/Model (started 0:07:12 ago; pid 795)
  Selecting TensorBoard with logdir Object Detection/Model (started 0:07:12 ago; port 6006, pid 795).

6. LOG_DIR = 'Object Detection/Model'
   get_ipython().system_raw(
       'tensorboard --logdir {} --host 0.0.0.0 --port 6006 &'
       .format(LOG_DIR)
   )
   # Install
   ! npm install -g localtunnel
   
   # Tunnel port 6006 (TensorBoard assumed running)
   get_ipython().system_raw('lt --port 6006 >> url.txt 2>&1 &')
   
   # Get url
   ! cat url.txt
   

/tools/node/bin/lt -> /tools/node/lib/node_modules/localtunnel/bin/lt.js
  + [email protected]
  updated 1 package in 1.601s

创建的url.txt文件为空

Answer 1

替换

!tensorboard --logdir='Object Detection/Model'

with

%tensorboard --logdir='Object Detection/Model'