prediction pipeline created

requirements.txt

@@ -15,4 +15,5 @@ scipy
 Flask
 Flask-Cors
 google-cloud-storage
+requests
 -e .

research/trails.ipynb

@@ -0,0 +1,239 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "21f6b22f-6f08-40a8-a161-636847421958",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "from PIL import Image"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "6bd4cd50-08c4-4603-a8a3-0eb9ac435666",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "os.chdir(\"../\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "c89c80e8-4c2c-4cb0-8146-90e25f1b65d1",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "'/home/lucky/Documents/projects/MyProject/smoke_detection'"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "pwd"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "3bd102a3-ce32-4e62-8812-1ba6ff2f978f",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "2024-01-04 22:17:37.107698: I external/local_tsl/tsl/cuda/cudart_stub.cc:31] Could not find cuda drivers on your machine, GPU will not be used.\n",
+      "2024-01-04 22:17:37.144805: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered\n",
+      "2024-01-04 22:17:37.144834: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered\n",
+      "2024-01-04 22:17:37.145880: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n",
+      "2024-01-04 22:17:37.151404: I external/local_tsl/tsl/cuda/cudart_stub.cc:31] Could not find cuda drivers on your machine, GPU will not be used.\n",
+      "2024-01-04 22:17:37.151900: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n",
+      "To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
+      "2024-01-04 22:17:37.915131: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT\n"
+     ]
+    }
+   ],
+   "source": [
+    "from cnnClassifier.pipeline.predict import PredictionPipeline"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "27f695ee-b870-40f4-a8f8-7f9f38d0497f",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "url = \"https://img.freepik.com/free-photo/young-man-smoking_144627-29295.jpg\""
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "6e4ede48-d531-4b0e-83ff-e74cf308c39e",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import requests"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "3e1e62d6-3bee-485d-bae3-9756159c78a9",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "response = requests.get(url)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "c661cd06-1a88-40c0-99b1-7f1505879269",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "image_bytes = response.content"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "141ceb01-2d89-48c9-b7ab-00e74e4f712b",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "[2024-01-04 22:17:39,290: INFO: common: yaml file: config/config.yaml loaded successfully]\n"
+     ]
+    }
+   ],
+   "source": [
+    "pp = PredictionPipeline()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "id": "ee4cbc61-2b25-451d-8a00-2cf2c1168ebc",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "image_bytes = pp.download_image(url2)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "e28cd59d-0637-41f0-9476-479563d2f037",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "img = pp.preprocess_image(image_bytes)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "fa2ae359-1aaa-4158-b822-180f1a1ecdeb",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(1, 224, 224, 3)"
+      ]
+     },
+     "execution_count": 19,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "img.shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "id": "7d135df6-7b7b-46ff-aca5-e71343a712ce",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Smoking with probability 94.80263590812683\n",
+      "Non-Smoking with probability 3.4822791814804077\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "{'SmokingProbability': 94.80263590812683,\n",
+       " 'NonSmokingProbability': 3.4822791814804077}"
+      ]
+     },
+     "execution_count": 20,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "pp.predict(img)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "56ae954a-a727-4b35-a99c-1075da6aae31",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "url2 = \"https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcQZ2FnaBLHoNCw4OM00db5ahJdvs_LXEo45OQ&usqp=CAU\""
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "38599440-8ddc-4666-b697-2598d34c64f5",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "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": "3.11.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

scores.json

@@ -1,4 +1,4 @@
 {
-    "loss": 0.12362710386514664,
-    "accuracy": 0.9508928656578064
+    "loss": 0.12056513875722885,
+    "accuracy": 0.9553571343421936
 }

src/cnnClassifier/pipeline/predict.py

@@ -0,0 +1,43 @@
+from PIL import Image
+from tensorflow.keras.models import load_model
+import numpy as np
+from cnnClassifier.config.configuration import ConfigurationManager
+from io import BytesIO
+import requests
+from cnnClassifier.constants import CONFIG_FILE_PATH
+from cnnClassifier.utils.common import read_yaml
+
+
+class PredictionPipeline:
+    def __init__(self):
+        self.config = read_yaml(CONFIG_FILE_PATH)
+        # load model
+        self.model = load_model(self.config.training.trained_model_path)
+        # self.params = read_yaml(PARAMS_FILE_PATH)
+
+    @staticmethod
+    def preprocess_image(image_bytes, img_shape=224):
+        """Will take the image bytes and return the processed image"""
+        image_data = BytesIO(image_bytes)
+        image = Image.open(image_data)
+        image = image.convert("RGB")
+        image = image.resize((img_shape, img_shape))
+        image_array = np.array(image)
+        img = np.expand_dims(image_array, axis=0)
+        return img
+
+    @staticmethod
+    def download_image(url):
+        """ will return image binary data"""
+        response = requests.get(url)
+        image_bytes = response.content
+        return image_bytes
+
+    def predict(self, img):
+        prediction = self.model.predict(img, verbose=0)
+        # predicted_label = np.argmax(prediction)
+        print(f"Smoking with probability {prediction[0][1] * 100}")
+        print(f"Non-Smoking with probability {prediction[0][0] * 100}")
+
+        return {"SmokingProbability": prediction[0][1] * 100,
+                "NonSmokingProbability": prediction[0][0]*100}