Skip to content

refactor sequential notebooks into independent notebooks for use-cases/computer_vision #8

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 7 commits into
base: main
Choose a base branch
from
Open

refactor sequential notebooks into independent notebooks for use-cases/computer_vision #8

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
058a7f2
refactor sequential notebooks into independent notebooks
atqy Apr 29, 2022
e1275a7
install opencv
atqy Apr 29, 2022
bd8b1c4
fix pip
atqy Apr 29, 2022
bba00ba
pip install command change
atqy Apr 29, 2022
a05d24c
Merge branch 'main' into atqy/computer-vision
atqy Apr 30, 2022
d144e64
reformat
atqy Apr 30, 2022
db4720a
Merge branch 'atqy/computer-vision' of https://github.com/atqy/amazon…
atqy Apr 30, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
710 changes: 0 additions & 710 deletions use-cases/computer_vision/1-metastases-detection-train-model.ipynb
View file
Open in desktop

This file was deleted.

296 changes: 286 additions & 10 deletions ...ion/4-metastases-detection-pipeline.ipynb → ...ision/metastases-detection-pipeline.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,8 +4,8 @@
"cell_type": "markdown",
"metadata": {},
"source": [
"# Computer Vision for Medical Imaging: Part 4. SageMaker Pipelines\n",
"This notebook is the final part of a 4-part series of techniques and services offer by SageMaker to build a model which predicts if an image of cells contains cancer. This notebook describes how to automate the ML workflow using SageMaker Pipelines."
"# Computer Vision for Medical Imaging - Pipeline Mode\n",
"This notebook showcases techniques and services offer by SageMaker to build a model which predicts if an image of cells contains cancer. This notebook describes how to automate the ML workflow using SageMaker Pipelines."
]
},
{
Expand Down Expand Up @@ -37,8 +37,7 @@
"metadata": {},
"outputs": [],
"source": [
"%store -r\n",
"%store"
"! pip install --upgrade sagemaker boto3"
]
},
{
Expand All @@ -48,6 +47,28 @@
"## Import Libraries"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"import pip\n",
"\n",
"\n",
"def import_or_install(package):\n",
" try:\n",
" __import__(package)\n",
" except ImportError:\n",
" ! pip install $package\n",
"\n",
"\n",
"required_packages = [\"sagemaker\", \"boto3\", \"h5py\", \"tqdm\", \"matplotlib\", \"opencv-python\"]\n",
"\n",
"for package in required_packages:\n",
" import_or_install(package)"
]
},
{
"cell_type": "code",
"execution_count": null,
Expand All @@ -59,6 +80,12 @@
"import numpy as np\n",
"import matplotlib.pyplot as plt\n",
"import cv2\n",
"import os\n",
"import zipfile\n",
"import h5py\n",
"import mxnet as mx\n",
"from datetime import datetime\n",
"from tqdm import tqdm\n",
"\n",
"from sagemaker.workflow.pipeline import Pipeline\n",
"from sagemaker.workflow.steps import CreateModelStep\n",
Expand Down Expand Up @@ -96,6 +123,208 @@
"bucket = sagemaker.Session().default_bucket()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Load Dataset"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# check if directory exists\n",
"if not os.path.isdir(\"data\"):\n",
" os.mkdir(\"data\")\n",
"\n",
"# download zip file from public s3 bucket\n",
"!wget -P data https://sagemaker-sample-files.s3.amazonaws.com/datasets/image/pcam/medical_images.zip"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"with zipfile.ZipFile(\"data/medical_images.zip\") as zf:\n",
" zf.extractall()\n",
"with open(\"data/camelyon16_tiles.h5\", \"rb\") as hf:\n",
" f = h5py.File(hf, \"r\")\n",
"\n",
" X = f[\"x\"][()]\n",
" y = f[\"y\"][()]\n",
"\n",
"print(\"Shape of X:\", X.shape)\n",
"print(\"Shape of y:\", y.shape)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# write to session s3 bucket\n",
"s3_client.upload_file(\"data/medical_images.zip\", bucket, f\"data/medical_images.zip\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# delete local copy\n",
"import os\n",
"\n",
"if os.path.exists(\"data/medical_images.zip\"):\n",
" os.remove(\"data/medical_images.zip\")\n",
"else:\n",
" print(\"The file does not exist\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## View Sample Images from Dataset"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def preview_images(X, y, n, cols):\n",
" sample_images = X[:n]\n",
" sample_labels = y[:n]\n",
"\n",
" rows = int(np.ceil(n / cols))\n",
" fig, axs = plt.subplots(rows, cols, figsize=(11.5, 7))\n",
"\n",
" for i, ax in enumerate(axs.flatten()):\n",
" image = sample_images[i]\n",
" label = sample_labels[i]\n",
" ax.imshow(image)\n",
" ax.axis(\"off\")\n",
" ax.set_title(f\"Label: {label}\")\n",
"\n",
" plt.tight_layout()\n",
"\n",
"\n",
"preview_images(X, y, 15, 5)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Shuffle and Split Dataset"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from sklearn.model_selection import train_test_split\n",
"\n",
"X_numpy = X[:]\n",
"y_numpy = y[:]\n",
"\n",
"X_train, X_test, y_train, y_test = train_test_split(\n",
" X_numpy, y_numpy, test_size=1000, random_state=0\n",
")\n",
"X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, test_size=2000, random_state=1)\n",
"\n",
"print(X_train.shape)\n",
"print(X_val.shape)\n",
"print(X_test.shape)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Convert Splits to RecordIO Format"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def write_to_recordio(X: np.ndarray, y: np.ndarray, prefix: str):\n",
" record = mx.recordio.MXIndexedRecordIO(idx_path=f\"{prefix}.idx\", uri=f\"{prefix}.rec\", flag=\"w\")\n",
" for idx, arr in enumerate(tqdm(X)):\n",
" header = mx.recordio.IRHeader(0, y[idx], idx, 0)\n",
" s = mx.recordio.pack_img(\n",
" header,\n",
" arr,\n",
" quality=95,\n",
" img_fmt=\".jpg\",\n",
" )\n",
" record.write_idx(idx, s)\n",
" record.close()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"write_to_recordio(X_train, y_train, prefix=\"data/train\")\n",
"write_to_recordio(X_val, y_val, prefix=\"data/val\")\n",
"write_to_recordio(X_test, y_test, prefix=\"data/test\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Upload Data Splits to S3"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"prefix = \"cv-metastasis\"\n",
"\n",
"try:\n",
" s3_client.create_bucket(\n",
" Bucket=bucket, ACL=\"private\", CreateBucketConfiguration={\"LocationConstraint\": region}\n",
" )\n",
" print(f\"Created S3 bucket: {bucket}\")\n",
"\n",
"except Exception as e:\n",
" if e.response[\"Error\"][\"Code\"] == \"BucketAlreadyOwnedByYou\":\n",
" print(f\"Using existing bucket: {bucket}\")\n",
" else:\n",
" raise (e)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"s3_client.upload_file(\"data/train.rec\", bucket, f\"{prefix}/data/train/train.rec\")\n",
"s3_client.upload_file(\"data/val.rec\", bucket, f\"{prefix}/data/val/val.rec\")\n",
"s3_client.upload_file(\"data/test.rec\", bucket, f\"{prefix}/data/test/test.rec\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
Expand All @@ -110,13 +339,15 @@
"outputs": [],
"source": [
"training_image = sagemaker.image_uris.retrieve(\"image-classification\", region)\n",
"num_training_samples = X_train.shape[0]\n",
"num_classes = len(np.unique(y_train))\n",
"\n",
"hyperparameters = {\n",
" \"num_layers\": 18,\n",
" \"use_pretrained_model\": 1,\n",
" \"augmentation_type\": \"crop_color_transform\",\n",
" \"image_shape\": \"3,96,96\",\n",
" \"num_classes\": 2,\n",
" \"num_classes\": num_classes,\n",
" \"num_training_samples\": num_training_samples,\n",
" \"mini_batch_size\": 64,\n",
" \"epochs\": 5,\n",
Expand Down Expand Up @@ -255,6 +486,8 @@
"metadata": {},
"outputs": [],
"source": [
"mpg_name = \"cv-metastasis-{}\".format(datetime.now().strftime(\"%Y-%m-%d-%H-%M-%S\"))\n",
"\n",
"model_approval_status = ParameterString(\n",
" name=\"ModelApprovalStatus\", default_value=\"PendingManualApproval\"\n",
")\n",
Expand Down Expand Up @@ -287,7 +520,7 @@
"source": [
"model = sagemaker.model.Model(\n",
" name=f\"{mpg_name}-pipline\",\n",
" image_uri=train_step.properties.AlgorithmSpecification.TrainingImage,\n",
" image_uri=training_image,\n",
" model_data=train_step.properties.ModelArtifacts.S3ModelArtifacts,\n",
" sagemaker_session=sagemaker_session,\n",
" role=sagemaker_role,\n",
Expand Down Expand Up @@ -315,6 +548,7 @@
" Filename=\"deploy_model.py\", Bucket=bucket, Key=f\"{prefix}/code/deploy_model.py\"\n",
")\n",
"deploy_model_script_uri = f\"s3://{bucket}/{prefix}/code/deploy_model.py\"\n",
"deploy_instance_type = \"ml.m4.xlarge\"\n",
"\n",
"deploy_model_processor = SKLearnProcessor(\n",
" framework_version=\"0.23-1\",\n",
Expand Down Expand Up @@ -355,7 +589,7 @@
"metadata": {},
"outputs": [],
"source": [
"pipeline_name = f\"{prefix}-pipeline\"\n",
"pipeline_name = \"{}-pipeline-{}\".format(prefix, datetime.now().strftime(\"%Y-%m-%d-%H-%M-%S\"))\n",
"\n",
"pipeline = Pipeline(\n",
" name=pipeline_name,\n",
Expand Down Expand Up @@ -419,7 +653,49 @@
"metadata": {},
"outputs": [],
"source": [
"best_model.sagemaker_session.delete_endpoint(mpg_name)"
"def delete_model_package_group(sm_client, package_group_name):\n",
" try:\n",
" model_versions = sm_client.list_model_packages(ModelPackageGroupName=package_group_name)\n",
"\n",
" except Exception as e:\n",
" print(\"{} \\n\".format(e))\n",
" return\n",
"\n",
" for model_version in model_versions[\"ModelPackageSummaryList\"]:\n",
" try:\n",
" sm_client.delete_model_package(ModelPackageName=model_version[\"ModelPackageArn\"])\n",
" except Exception as e:\n",
" print(\"{} \\n\".format(e))\n",
" time.sleep(0.5) # Ensure requests aren't throttled\n",
"\n",
" try:\n",
" sm_client.delete_model_package_group(ModelPackageGroupName=package_group_name)\n",
" print(\"{} model package group deleted\".format(package_group_name))\n",
" except Exception as e:\n",
" print(\"{} \\n\".format(e))\n",
" return\n",
"\n",
"\n",
"def delete_sagemaker_pipeline(sm_client, pipeline_name):\n",
" try:\n",
" sm_client.delete_pipeline(\n",
" PipelineName=pipeline_name,\n",
" )\n",
" print(\"{} pipeline deleted\".format(pipeline_name))\n",
" except Exception as e:\n",
" print(\"{} \\n\".format(e))\n",
" return"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"client = sagemaker.Session().sagemaker_client\n",
"delete_model_package_group(client, mpg_name)\n",
"delete_sagemaker_pipeline(client, pipeline_name)"
]
},
{
Expand All @@ -433,9 +709,9 @@
"metadata": {
"instance_type": "ml.t3.medium",
"kernelspec": {
"display_name": "conda_python3",
"display_name": "conda_mxnet_p36",
"language": "python",
"name": "conda_python3"
"name": "conda_mxnet_p36"
},
"language_info": {
"codemirror_mode": {
Expand Down
Loading