dataset_extraction.py

data = {}
p = r"E:\Documents\My Projects\Instagram Dashboard\Model Development\flower_features.pkl"

# lop through each image in the dataset
for flower in flowers:
    # try to extract the features and update the dictionary
    try:
        feat = extract_features(flower,model)
        data[flower] = feat
    # if something fails, save the extracted features as a pickle file (optional)
    except:
        with open(p,'wb') as file:
            pickle.dump(data,file)
          
 
# get a list of the filenames
filenames = np.array(list(data.keys()))

# get a list of just the features
feat = np.array(list(data.values()))
feat.shape
(210, 1, 4096)

# reshape so that there are 210 samples of 4096 vectors
feat = feat.reshape(-1,4096)
feat.shape
(210, 4096)

# get the unique labels (from the flower_labels.csv)
df = pd.read_csv('flower_labels.csv')
label = df['label'].tolist()
unique_labels = list(set(label))