explaining image classifiers initial commit 🎉

.gitignore

@@ -1,4 +1,21 @@
-.idea/*
-**/data
-.env/*
+# Notebook data files
+notebooks/*.pkl
+notebooks/*.tsv
+# Jupyter checkpoints
 **/.ipynb_checkpoints
+
+# PyCharm and mac and azure ml and VS Code files
+.idea
+.DS_Store
+.azureml
+.vscode
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# Environments
+**/data
+**/dataset
+**/.env

02_data_science_toolkit/03_explaining_image_classifier_predictions/01_coco_res_net/coco_loader.py

@@ -0,0 +1,22 @@
+from skimage.io import imread
+from skimage.transform import resize
+import numpy as np
+
+class CocoLoader:
+
+    IMG_WIDTH = 224
+    IMG_HEIGHT = 224
+    IMG_CHANNEL = 3
+
+    def preprocess_image(self, path: str):
+        x = imread(path)
+        x = resize(x, (CocoLoader.IMG_WIDTH, CocoLoader.IMG_HEIGHT)) * 255
+        return x
+
+    def load_sample(self, size: int):
+        with open("./dataset/5k.txt", "r") as f:
+            images_paths = f.read().split('\n')
+            dataset = np.ndarray(shape=(size, CocoLoader.IMG_WIDTH, CocoLoader.IMG_HEIGHT, CocoLoader.IMG_CHANNEL), dtype=np.float32)
+            for index, images_path in enumerate(images_paths[:size]):
+                dataset[index] = self.preprocess_image(images_path)
+        return dataset

02_data_science_toolkit/03_explaining_image_classifier_predictions/01_coco_res_net/get_coco_dataset_sample.sh

@@ -0,0 +1,28 @@
+#!/bin/bash
+
+# CREDIT: https://github.com/pjreddie/darknet/tree/master/scripts/get_coco_dataset.sh
+
+mkdir dataset
+cd dataset
+
+mkdir coco
+cd coco
+
+mkdir images
+cd images
+
+# Download Images
+wget -c https://pjreddie.com/media/files/val2014.zip
+
+# Unzip
+unzip -q val2014.zip
+
+cd ..
+cd ..
+
+wget -c https://pjreddie.com/media/files/coco/5k.part
+wget -c https://pjreddie.com/media/files/coco/labels.tgz
+tar xzf labels.tgz
+
+# Set Up Image Lists
+paste <(awk "{print \"$PWD/coco\"}" <5k.part) 5k.part | tr -d '\t' > 5k.txt