step3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "d6da830e-0354-4b79-aa0c-681e876a3eb5",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-11-30 13:29:30.309486: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2023-11-30 13:29:31.377988: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer.so.7'; dlerror: libnvinfer.so.7: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/nvidia/lib:/usr/local/nvidia/lib64\n",
      "2023-11-30 13:29:31.378069: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer_plugin.so.7'; dlerror: libnvinfer_plugin.so.7: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/nvidia/lib:/usr/local/nvidia/lib64\n",
      "2023-11-30 13:29:31.378078: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras import Sequential\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import accuracy_score\n",
    "from datetime import datetime\n",
    "import tf2onnx\n",
    "import onnx\n",
    "import onnxruntime as ort\n",
    "import os\n",
    "import boto3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "7bc0f017-d615-4b7f-a719-08c076c8b4b3",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "(X_train,y_train),(X_test,y_test) = keras.datasets.mnist.load_data()\n",
    "\n",
    "X_train = X_train/255\n",
    "X_test = X_test/255"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "395b95ce-f97e-4927-ae48-df361f3f06c0",
   "metadata": {},
   "source": [
    "# Training\n",
    "\n",
    "Let's train a an alternative, a convoluted neural network:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "b2ce7e47-dc57-4156-9808-bf5de1c02675",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d (Conv2D)             (None, 26, 26, 32)        320       \n",
      "                                                                 \n",
      " max_pooling2d (MaxPooling2D  (None, 13, 13, 32)       0         \n",
      " )                                                               \n",
      "                                                                 \n",
      " flatten (Flatten)           (None, 5408)              0         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 64)                346176    \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 64)                0         \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 10)                650       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 347,146\n",
      "Trainable params: 347,146\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "\n",
    "model.add(tf.keras.layers.Conv2D(filters=32, kernel_size=(3, 3), input_shape=(28, 28, 1)))\n",
    "model.add(tf.keras.layers.MaxPooling2D(pool_size=(2, 2), strides=2))\n",
    "model.add(tf.keras.layers.Flatten())\n",
    "model.add(tf.keras.layers.Dense(units=64, activation=tf.nn.relu))\n",
    "model.add(tf.keras.layers.Dropout(rate=0.2))\n",
    "model.add(tf.keras.layers.Dense(10, activation=tf.nn.softmax))\n",
    "\n",
    "model.compile(optimizer='adam',\n",
    "              loss='sparse_categorical_crossentropy',\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "9e5fbbf9-fa00-49a4-b44a-fa0e876db003",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/3\n",
      "1875/1875 [==============================] - 28s 14ms/step - loss: 0.2460 - accuracy: 0.9266\n",
      "Epoch 2/3\n",
      "1875/1875 [==============================] - 27s 14ms/step - loss: 0.1094 - accuracy: 0.9665\n",
      "Epoch 3/3\n",
      "1875/1875 [==============================] - 27s 14ms/step - loss: 0.0810 - accuracy: 0.9747\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(X_train,y_train,epochs=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "81818fe7-cfdf-4e1d-a539-13077da077cc",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 4ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.9794"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_prob = model.predict(X_test)\n",
    "y_pred = y_prob.argmax(axis=1)\n",
    "accuracy_score(y_test,y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4e0cb805-119c-4f4d-b180-6b11ea17852f",
   "metadata": {},
   "source": [
    "# Demo\n",
    "\n",
    "Let's try out the nn with a small demo, given the following:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "f62e879f-c7c3-432e-9592-24f767923589",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(X_test[4], cmap=\"Greys\")\n",
    "plt.title('X_test[4]:')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4f2dd675-8171-4bcb-a523-130027e625c9",
   "metadata": {},
   "source": [
    "the model predicts it is:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "0e4427b4-58c8-4712-a72c-277e7de1ed03",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 18ms/step\n",
      "prediction:  4\n"
     ]
    }
   ],
   "source": [
    "prediction = model.predict(X_test[4].reshape(1,28,28)).argmax(axis=1)[0]\n",
    "print('prediction: ', prediction)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec17cb9b-6ff3-4b83-9f65-b27264b1552b",
   "metadata": {},
   "source": [
    "# Model Save"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "2dff55b1-8446-4ee7-9e11-15d378e3875e",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "input_signature = [tf.TensorSpec([1, 28, 28], tf.double, name='x')]\n",
    "onnx_model, _ = tf2onnx.convert.from_keras(model, input_signature, opset=12)\n",
    "onnx.save(onnx_model, \"model2.onnx\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "384b077d-5a1c-4f0f-a48d-41fbc96d0bf4",
   "metadata": {},
   "source": [
    "# Demo using onnx file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "f6d61319-9860-4040-885d-64e071057340",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['dense_1']\n",
      "prediction:  4\n"
     ]
    }
   ],
   "source": [
    "model_onnx = onnx.load('model2.onnx')\n",
    "output = [node.name for node in model_onnx.graph.output]\n",
    "print(output)\n",
    "\n",
    "sess = ort.InferenceSession(\"model2.onnx\")\n",
    "results_ort = sess.run(['dense_1'], {'x': X_test[4].reshape(1,28,28)})\n",
    "prediction = results_ort[0].argmax(axis=1)[0]\n",
    "print('prediction: ', prediction)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9fcdbf7e-2e4f-4dee-81d6-b4993a0cd1d4",
   "metadata": {},
   "source": [
    "# Store onnx on bucket"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "2a068fdc-cd9e-4dc7-966c-eb29d8dc8004",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Will be using: mnist:v2.nb20231130133057 in the remainder of this task\n",
      "ls:\n",
      "artifacts/test-matteo-storesomething-11c33/train-model/outgoingfile.tgz\n",
      "artifacts/test-matteo-storesomething-11c33/train-model/outgoingfile2.tgz\n",
      "artifacts/test-matteo-storesomething-11c33/validate-model/Output.tgz\n",
      "experiment1/mnist.onnx\n",
      "pipelines/91c05d13-7a85-4133-844e-889c3a00999c\n",
      "pipelines/db1f4601-8c10-4379-8a22-09a49faa66e5\n",
      "v1.nb20231130105804/mnist.onnx\n",
      "v1.nb20231130105849/mnist.onnx\n",
      "v1.nb20231130110240/mnist.onnx\n",
      "v1.nb20231130110352/mnist.onnx\n",
      "v1.nb20231130132831/mnist.onnx\n",
      "v2.nb20231130132332/mnist.onnx\n",
      "v2.nb20231130133057/mnist.onnx\n",
      "v20231128110416/mnist.onnx\n"
     ]
    }
   ],
   "source": [
    "registeredmodel_name = \"mnist\"\n",
    "version_name = \"v2.nb\"+datetime.now().strftime(\"%Y%m%d%H%M%S\")\n",
    "print(f\"Will be using: {registeredmodel_name}:{version_name} in the remainder of this task\")\n",
    "\n",
    "s3 = boto3.resource(\n",
    "    service_name='s3',\n",
    "    region_name=os.environ['AWS_DEFAULT_REGION'],\n",
    "    aws_access_key_id=os.environ['AWS_ACCESS_KEY_ID'],\n",
    "    aws_secret_access_key=os.environ['AWS_SECRET_ACCESS_KEY'],\n",
    "    use_ssl=False,\n",
    "    endpoint_url=os.environ['AWS_S3_ENDPOINT'],\n",
    "    verify=False\n",
    ")\n",
    "\n",
    "bucket_name = os.environ['AWS_S3_BUCKET']\n",
    "odh_secret_name = f'aws-connection-{bucket_name}'\n",
    "in_bucket_path = version_name\n",
    "in_bucket_target = f'{in_bucket_path}/mnist.onnx'\n",
    "full_bucket_target = f's3://{bucket_name}/{in_bucket_target}'\n",
    "\n",
    "my_bucket = s3.Bucket(bucket_name)\n",
    "my_bucket.upload_file(\"model2.onnx\", in_bucket_target)\n",
    "\n",
    "print(\"ls:\")\n",
    "for obj in my_bucket.objects.filter():\n",
    "    print(obj.key)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "33512d06-ce8c-46ab-96f6-f0aaa1875194",
   "metadata": {},
   "source": [
    "# Register model on Model Registry"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "6ae4b66b-707f-4eb6-b693-568d6ec7072b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RegisteredModel:\n",
      "RegisteredModel(id='2', description='', external_id='', create_time_since_epoch=1701169456901, last_update_time_since_epoch=1701169456901, name='mnist')\n",
      "ModelVersion:\n",
      "ModelVersion(name='v2.nb20231130133057', id='14', description='demo20231121 e2e MNIST', external_id='', create_time_since_epoch=1701351081485, last_update_time_since_epoch=1701351081485, model_name='mnist', version='v2.nb20231130133057', author='author', artifacts=NOTHING, tags=[], metadata={}, _registered_model_id=NOTHING)\n",
      "ModelArtifact:\n",
      "ModelArtifact(id='8', description='', external_id='', create_time_since_epoch=1701351082404, last_update_time_since_epoch=1701351082404, name='mnist', uri='s3://mybucket/v2.nb20231130133057/mnist.onnx', state=<ArtifactState.UNKNOWN: 0>, model_format_name='onnx', model_format_version='1', storage_key='aws-connection-mybucket', storage_path='v2.nb20231130133057', service_account_name='', _model_version_id=NOTHING)\n"
     ]
    }
   ],
   "source": [
    "from model_registry import ModelRegistry\n",
    "\n",
    "registry = ModelRegistry(server_address=\"modelregistry-sample\", port=9090, author=\"author\")\n",
    "\n",
    "rm = registry.register_model(registeredmodel_name,\n",
    "                                full_bucket_target,\n",
    "                                model_format_name=\"onnx\",\n",
    "                                model_format_version=\"1\",\n",
    "                                storage_key=odh_secret_name,\n",
    "                                storage_path=in_bucket_path,\n",
    "                                version=version_name,\n",
    "                                description=\"demo20231121 e2e MNIST\",\n",
    "                                )\n",
    "print(\"RegisteredModel:\")\n",
    "print(registry.get_registered_model(registeredmodel_name))\n",
    "print(\"ModelVersion:\")\n",
    "print(registry.get_model_version(registeredmodel_name, version_name))\n",
    "print(\"ModelArtifact:\")\n",
    "print(registry.get_model_artifact(registeredmodel_name, version_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1442f3ca-4206-487b-bae6-f6d00a036368",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}