Add comments about tensorboard and fix kernel for all notebooks.

NASA-IMPACT · May 25, 2021 · 05ebf20 · 05ebf20
1 parent 1cb0b64
commit 05ebf20
Show file tree

Hide file tree

Showing 4 changed files with 26 additions and 14 deletions.
diff --git a/chapter-1/imagelabeler_details.ipynb b/chapter-1/imagelabeler_details.ipynb
@@ -69,14 +69,14 @@
  ],
  "metadata": {
  "kernelspec": {
- "display_name": "Python 3",
+ "display_name": "conda_tensorflow2_p36",
  "language": "python",
- "name": "python3"
+ "name": "conda_tensorflow2_p36"
  },
  "language_info": {
  "codemirror_mode": {
  "name": "ipython",
- "version": 3
+ "version": "3.6.13"
  },
  "file_extension": ".py",
  "mimetype": "text/x-python",

diff --git a/chapter-2/image_labeler_data_download.ipynb b/chapter-2/image_labeler_data_download.ipynb
@@ -1102,13 +1102,25 @@
  ]
  },
  {
+ "cell_type": "markdown",
+ "metadata": {},
  "source": [
  "# Note the trained model name in the DESTINATION_BUCKET dashboard\n",
  "\n",
- "This will be used in the next chapter to deploy and infer from the model."
- ],
- "cell_type": "markdown",
- "metadata": {}
+ "This will be used in the next chapter to deploy and infer from the model.\n",
+ "\n",
+ "\n",
+ "## Attach a Tensorboard session\n",
+ "\n",
+ "Tensorboard in an interactive model training visualization tool. This lets us monitor the training/validation losses over time and gives an idea of how good the model is training. (Since we are passing the logs to s3, we can interact with the logs directly from our local setup.)\n",
+ "Here's how you'd connect your tensorboard to your s3 bucket.\n",
+ "1. install tensorboard version 2.4.1 using `pip install tensorboard==2.4.1`\n",
+ "2. run `AWS_REGION=<the region you selected> tensorboard --logdir s3:https://<your account number>-model-bucket/tensorboard_logs/`\n",
+ "3. Navigate to localhost:6006 to view the model metrics, model graph timeline and more.\n",
+ "\n",
+ "\n",
+ "You could also try running it via this very notebook using the following cell. (Will be useful in chapter-3)"
+ ]
  },
  {
  "cell_type": "code",
@@ -1118,7 +1130,7 @@
  "source": [
  "LOG_FOLDER = \"tensorboard_logs\"\n",
  "TENSORFLOW_LOGS_PATH = f\"s3:https://{BUCKET_NAME}/{LOG_FOLDER}\"\n",
- "\n",
+ "# The corresponding tensorboard can be accessed using `https://<your-notebook-instance-name>/proxy/6006/` the https and the trailing spaces are very important.\n",
  "!AWS_REGION=us-east-1 tensorboard --logdir {TENSORFLOW_LOGS_PATH}"
  ]
  }

diff --git a/chapter-3/model_training.ipynb b/chapter-3/model_training.ipynb
@@ -245,14 +245,14 @@
  ],
  "metadata": {
  "kernelspec": {
- "display_name": "Python 3",
+ "display_name": "conda_tensorflow2_p36",
  "language": "python",
- "name": "python3"
+ "name": "conda_tensorflow2_p36"
  },
  "language_info": {
  "codemirror_mode": {
  "name": "ipython",
- "version": 3
+ "version": "3.6.13"
  },
  "file_extension": ".py",
  "mimetype": "text/x-python",

diff --git a/chapter-4/model_deployment_and_test.ipynb b/chapter-4/model_deployment_and_test.ipynb
@@ -211,14 +211,14 @@
  ],
  "metadata": {
  "kernelspec": {
- "display_name": "Python 3",
+ "display_name": "conda_tensorflow2_p36",
  "language": "python",
- "name": "python3"
+ "name": "conda_tensorflow2_p36"
  },
  "language_info": {
  "codemirror_mode": {
  "name": "ipython",
- "version": 3
+ "version": "3.6.13"
  },
  "file_extension": ".py",
  "mimetype": "text/x-python",