{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"## (In Progress)","metadata":{"id":"b85cf361"}},{"cell_type":"markdown","source":"### Note: The code is setup in a way that you can easily run the notebook on colab with minimal changes\n### TFRecords Dataset: https://www.kaggle.com/ks2019/happywhale-tfrecords-v1\n### Code for generating TFRecords: https://www.kaggle.com/ks2019/happywhale-tfrecords","metadata":{"id":"061c9adb"}},{"cell_type":"code","source":"import os\nIS_COLAB = not os.path.exists('/kaggle/input')\nprint(IS_COLAB) ","metadata":{"id":"dc57b06f","outputId":"30d7413b-3e5b-46e8-8107-a7469156ea4a","execution":{"iopub.status.busy":"2022-02-22T14:11:24.350592Z","iopub.execute_input":"2022-02-22T14:11:24.351143Z","iopub.status.idle":"2022-02-22T14:11:24.379571Z","shell.execute_reply.started":"2022-02-22T14:11:24.351047Z","shell.execute_reply":"2022-02-22T14:11:24.378854Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#!pip install tfimm\n#!pip install timm\n#import tfimm\n#print(tfimm.list_models(pretrained=\"timm\"))","metadata":{"id":"tG-UfOvDa1Qj","outputId":"46b2de7a-9887-4e9d-91cf-60892f2193e1","execution":{"iopub.status.busy":"2022-02-22T14:11:24.380918Z","iopub.execute_input":"2022-02-22T14:11:24.381509Z","iopub.status.idle":"2022-02-22T14:11:24.387914Z","shell.execute_reply.started":"2022-02-22T14:11:24.381478Z","shell.execute_reply":"2022-02-22T14:11:24.387056Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import tensorflow as tf\ntry:\n    # TPU detection. No parameters necessary if TPU_NAME environment variable is\n    # set: this is always the case on Kaggle.\n    tpu = tf.distribute.cluster_resolver.TPUClusterResolver()\n    print('Running on TPU ', tpu.master())\nexcept ValueError:\n    tpu = None\n\nif tpu:\n    tf.config.experimental_connect_to_cluster(tpu)\n    tf.tpu.experimental.initialize_tpu_system(tpu)\n    strategy = tf.distribute.TPUStrategy(tpu)\nelse:\n    # Default distribution strategy in Tensorflow. Works on CPU and single GPU.\n    strategy = tf.distribute.get_strategy()\n\nAUTO = tf.data.experimental.AUTOTUNE\nprint(\"REPLICAS: \", strategy.num_replicas_in_sync)","metadata":{"id":"562beb24","outputId":"c93079ce-86cb-44d1-dda7-1e8336001330","execution":{"iopub.status.busy":"2022-02-22T14:11:24.389107Z","iopub.execute_input":"2022-02-22T14:11:24.389475Z","iopub.status.idle":"2022-02-22T14:11:35.979192Z","shell.execute_reply.started":"2022-02-22T14:11:24.389444Z","shell.execute_reply":"2022-02-22T14:11:35.977665Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"if IS_COLAB:\n  from google.colab import drive\n  drive.mount('/content/drive')\nelse:\n  from kaggle_datasets import KaggleDatasets","metadata":{"id":"8f987118","outputId":"20767180-e81d-4069-e1ce-51065c7c26d3","execution":{"iopub.status.busy":"2022-02-22T14:11:35.982307Z","iopub.execute_input":"2022-02-22T14:11:35.982695Z","iopub.status.idle":"2022-02-22T14:11:35.990526Z","shell.execute_reply.started":"2022-02-22T14:11:35.982647Z","shell.execute_reply":"2022-02-22T14:11:35.989317Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"!pip install -q efficientnet\n!pip install tensorflow_addons\nimport re\nimport os\nimport numpy as np\nimport pandas as pd\nimport random\nimport math\nimport tensorflow as tf\nimport efficientnet.tfkeras as efn\nfrom sklearn import metrics\nfrom sklearn.model_selection import KFold, train_test_split\nfrom tensorflow.keras import backend as K\nimport tensorflow_addons as tfa\nfrom tqdm.auto import tqdm\nimport matplotlib.pyplot as plt\nimport pickle\nimport json\nimport tensorflow_hub as tfhub\nfrom datetime import datetime","metadata":{"id":"3e7115e8","outputId":"a8aacde9-8ac6-4efa-903d-bc4b5b8d9941","execution":{"iopub.status.busy":"2022-02-22T14:11:35.994439Z","iopub.execute_input":"2022-02-22T14:11:35.994930Z","iopub.status.idle":"2022-02-22T14:11:55.959269Z","shell.execute_reply.started":"2022-02-22T14:11:35.994886Z","shell.execute_reply":"2022-02-22T14:11:55.958040Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Config","metadata":{"id":"acf326ac"}},{"cell_type":"code","source":"save_dir = '.'\nEXPERIMENT = 0\nrun_ts = datetime.now().strftime('%Y%m%d-%H%M%S')\nprint(run_ts)\nif IS_COLAB:\n    save_dir = f'/content/drive/MyDrive/Models/'\n    #!mkdir -p {save_dir}","metadata":{"id":"8d270120","outputId":"cf8a3012-2877-4fbf-e6dc-6d0897e0d9ee","execution":{"iopub.status.busy":"2022-02-22T14:11:55.961123Z","iopub.execute_input":"2022-02-22T14:11:55.961423Z","iopub.status.idle":"2022-02-22T14:11:55.967515Z","shell.execute_reply.started":"2022-02-22T14:11:55.961382Z","shell.execute_reply":"2022-02-22T14:11:55.966508Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class config:\n    \n    \n    SEED = 42\n    FOLD_TO_RUN = 4\n    FOLDS = 5\n    DEBUG = False\n    EVALUATE = True\n    RESUME = True\n    RESUME_EPOCH = 9\n    \n    \n    ### Dataset\n    BATCH_SIZE = 4 * strategy.num_replicas_in_sync\n    IMAGE_SIZE = 768\n    N_CLASSES = 15587\n    \n    ### Model\n    model_type = 'effnetv1' #'effnetv1'  \n    EFF_NET = 7\n    EFF_NETV2 = 's-21k-ft1k'\n    FREEZE_BATCH_NORM = False\n    head = 'arcface' \n    EPOCHS = 21\n    LR = 0.001\n    message='baseline'\n    \n    ### Augmentations\n    CUTOUT = False\n    \n    ### Save-Directory\n    save_dir = save_dir\n    \n    ### Inference\n    KNN = 100\n    \ndef count_data_items(filenames):\n    n = [int(re.compile(r\"-([0-9]*)\\.\").search(filename).group(1)) \n         for filename in filenames]\n    return np.sum(n)\n\n# Function to seed everything\ndef seed_everything(seed):\n    random.seed(seed)\n    np.random.seed(seed)\n    os.environ['PYTHONHASHSEED'] = str(seed)\n    tf.random.set_seed(seed)\n    \ndef is_interactive():\n    return 'runtime'    in get_ipython().config.IPKernelApp.connection_file\nIS_INTERACTIVE = is_interactive()\nprint(IS_INTERACTIVE)","metadata":{"id":"0a9aa72d","outputId":"a666bf64-b580-436b-fd91-26e952f09f24","execution":{"iopub.status.busy":"2022-02-22T14:11:55.968713Z","iopub.execute_input":"2022-02-22T14:11:55.968955Z","iopub.status.idle":"2022-02-22T14:11:55.983000Z","shell.execute_reply.started":"2022-02-22T14:11:55.968919Z","shell.execute_reply":"2022-02-22T14:11:55.982284Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"MODEL_NAME = None\nif config.model_type == 'effnetv1':\n    MODEL_NAME = f'effnetv1_b{config.EFF_NET}'\nelif config.model_type == 'effnetv2':\n    MODEL_NAME = f'effnetv2_{config.EFF_NETV2}'\n\nconfig.MODEL_NAME = MODEL_NAME\nprint(MODEL_NAME)","metadata":{"id":"ffa3e85d","outputId":"c77fa6f1-6d1b-44d8-807e-439d18297671","execution":{"iopub.status.busy":"2022-02-22T14:11:55.984266Z","iopub.execute_input":"2022-02-22T14:11:55.984664Z","iopub.status.idle":"2022-02-22T14:11:55.997732Z","shell.execute_reply.started":"2022-02-22T14:11:55.984633Z","shell.execute_reply":"2022-02-22T14:11:55.997093Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with open(config.save_dir+'/config.json', 'w') as fp:\n    json.dump({x:dict(config.__dict__)[x] for x in dict(config.__dict__) if not x.startswith('_')}, fp)","metadata":{"id":"f23edd6d","execution":{"iopub.status.busy":"2022-02-22T14:11:55.999470Z","iopub.execute_input":"2022-02-22T14:11:55.999962Z","iopub.status.idle":"2022-02-22T14:11:56.010171Z","shell.execute_reply.started":"2022-02-22T14:11:55.999920Z","shell.execute_reply":"2022-02-22T14:11:56.009386Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"GCS_PATH = 'gs://kds-2f25b435592b59d6a2e92f82f0316665f6b69e4768d1296342746e85'  # Get GCS Path from kaggle notebook if GCS Path is expired\nif not IS_COLAB:\n    GCS_DS_PATH=KaggleDatasets().get_gcs_path('happywhale-tfrecords-bb')\n    \ntrain_files = np.sort(np.array(tf.io.gfile.glob(GCS_PATH + '/happywhale-2022-train*.tfrec')))\ntest_files = np.sort(np.array(tf.io.gfile.glob(GCS_PATH + '/happywhale-2022-test*.tfrec')))\nprint(GCS_PATH)\nprint(len(train_files),len(test_files),count_data_items(train_files),count_data_items(test_files))","metadata":{"id":"fd3e4646","outputId":"f02f2dd6-98e0-445e-cf8b-a8ce83c62e24","execution":{"iopub.status.busy":"2022-02-22T14:11:56.013158Z","iopub.execute_input":"2022-02-22T14:11:56.013708Z","iopub.status.idle":"2022-02-22T14:11:57.081409Z","shell.execute_reply.started":"2022-02-22T14:11:56.013667Z","shell.execute_reply":"2022-02-22T14:11:57.080697Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Data","metadata":{"id":"a2f96447"}},{"cell_type":"code","source":"def arcface_format(posting_id, image, label_group, matches):\n    return posting_id, {'inp1': image, 'inp2': label_group}, label_group, matches\n\ndef arcface_inference_format(posting_id, image, label_group, matches):\n    return image,posting_id\n\ndef arcface_eval_format(posting_id, image, label_group, matches):\n    return image,label_group\n\n# Data augmentation function\ndef data_augment(posting_id, image, label_group, matches):\n\n    ### CUTOUT\n    if tf.random.uniform([])>0.5 and config.CUTOUT:\n      N_CUTOUT = 6\n      for cutouts in range(N_CUTOUT):\n        if tf.random.uniform([])>0.5:\n           DIM = config.IMAGE_SIZE\n           CUTOUT_LENGTH = DIM//8\n           x1 = tf.cast( tf.random.uniform([],0,DIM-CUTOUT_LENGTH),tf.int32)\n           x2 = tf.cast( tf.random.uniform([],0,DIM-CUTOUT_LENGTH),tf.int32)\n           filter_ = tf.concat([tf.zeros((x1,CUTOUT_LENGTH)),tf.ones((CUTOUT_LENGTH,CUTOUT_LENGTH)),tf.zeros((DIM-x1-CUTOUT_LENGTH,CUTOUT_LENGTH))],axis=0)\n           filter_ = tf.concat([tf.zeros((DIM,x2)),filter_,tf.zeros((DIM,DIM-x2-CUTOUT_LENGTH))],axis=1)\n           cutout = tf.reshape(1-filter_,(DIM,DIM,1))\n           image = cutout*image\n\n    image = tf.image.random_flip_left_right(image)\n    # image = tf.image.random_flip_up_down(image)\n    image = tf.image.random_hue(image, 0.01)\n    image = tf.image.random_saturation(image, 0.70, 1.30)\n    image = tf.image.random_contrast(image, 0.80, 1.20)\n    image = tf.image.random_brightness(image, 0.10)\n    return posting_id, image, label_group, matches\n\n# Function to decode our images\n# Updated to include crops.\ndef decode_image(image_data, box):\n    if box is not None and box[0] != -1:\n        left, top, right, bottom = box[0], box[1], box[2], box[3]\n        bbs = tf.convert_to_tensor([top, left, bottom - top, right - left])\n        image = tf.io.decode_and_crop_jpeg(image_data, bbs, channels=3)\n    else:\n        image = tf.image.decode_jpeg(image_data, channels = 3)\n\n    image = tf.image.resize(image, [config.IMAGE_SIZE,config.IMAGE_SIZE])\n    image = tf.cast(image, tf.float32) / 255.0\n    return image\n\n# This function parse our images and also get the target variable\ndef read_labeled_tfrecord(example):\n    LABELED_TFREC_FORMAT = {\n        \"image_name\": tf.io.FixedLenFeature([], tf.string),\n        \"image\": tf.io.FixedLenFeature([], tf.string),\n        \"target\": tf.io.FixedLenFeature([], tf.int64),\n        'detic_box': tf.io.FixedLenFeature([4], tf.int64),\n         # 'yolov5_box': tf.io.FixedLenFeature([4], tf.int64),\n    }\n\n    example = tf.io.parse_single_example(example, LABELED_TFREC_FORMAT)\n    posting_id = example['image_name']\n    bb = tf.cast(example['detic_box'], tf.int32)\n    image = decode_image(example['image'], bb)\n#     label_group = tf.one_hot(tf.cast(example['label_group'], tf.int32), depth = N_CLASSES)\n    label_group = tf.cast(example['target'], tf.int32)\n#     matches = example['matches']\n    matches = 1\n    return posting_id, image, label_group, matches\n\n# This function loads TF Records and parse them into tensors\ndef load_dataset(filenames, ordered = False):\n    \n    ignore_order = tf.data.Options()\n    if not ordered:\n        ignore_order.experimental_deterministic = False \n        \n    dataset = tf.data.TFRecordDataset(filenames, num_parallel_reads = AUTO)\n#     dataset = dataset.cache()\n    dataset = dataset.with_options(ignore_order)\n    dataset = dataset.map(read_labeled_tfrecord, num_parallel_calls = AUTO) \n    return dataset\n\n# This function is to get our training tensors\ndef get_training_dataset(filenames):\n    dataset = load_dataset(filenames, ordered = False)\n    dataset = dataset.map(data_augment, num_parallel_calls = AUTO)\n    dataset = dataset.map(arcface_format, num_parallel_calls = AUTO)\n    dataset = dataset.map(lambda posting_id, image, label_group, matches: (image, label_group))\n    dataset = dataset.repeat()\n    dataset = dataset.shuffle(2048)\n    dataset = dataset.batch(config.BATCH_SIZE)\n    dataset = dataset.prefetch(AUTO)\n    return dataset\n\n# This function is to get our training tensors\ndef get_val_dataset(filenames):\n    dataset = load_dataset(filenames, ordered = True)\n    dataset = dataset.map(data_augment, num_parallel_calls = AUTO)\n    dataset = dataset.map(arcface_format, num_parallel_calls = AUTO)\n    dataset = dataset.map(lambda posting_id, image, label_group, matches: (image, label_group))\n    dataset = dataset.batch(config.BATCH_SIZE)\n    dataset = dataset.prefetch(AUTO)\n    return dataset\n\n# This function is to get our training tensors\ndef get_eval_dataset(filenames, get_targets = True):\n    dataset = load_dataset(filenames, ordered = True)\n    dataset = dataset.map(data_augment, num_parallel_calls = AUTO)\n    dataset = dataset.map(arcface_eval_format, num_parallel_calls = AUTO)\n    if not get_targets:\n        dataset = dataset.map(lambda image, target: image)\n    dataset = dataset.batch(config.BATCH_SIZE)\n    dataset = dataset.prefetch(AUTO)\n    return dataset\n\n# This function is to get our training tensors\ndef get_test_dataset(filenames, get_names = True):\n    dataset = load_dataset(filenames, ordered = True)\n    dataset = dataset.map(data_augment, num_parallel_calls = AUTO)\n    dataset = dataset.map(arcface_inference_format, num_parallel_calls = AUTO)\n    if not get_names:\n        dataset = dataset.map(lambda image, posting_id: image)\n    dataset = dataset.batch(config.BATCH_SIZE)\n    dataset = dataset.prefetch(AUTO)\n    return dataset","metadata":{"id":"65cf2a82","execution":{"iopub.status.busy":"2022-02-22T14:11:57.082551Z","iopub.execute_input":"2022-02-22T14:11:57.083309Z","iopub.status.idle":"2022-02-22T14:11:57.117961Z","shell.execute_reply.started":"2022-02-22T14:11:57.083263Z","shell.execute_reply":"2022-02-22T14:11:57.117240Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"row = 10; col = 8;\nrow = min(row,config.BATCH_SIZE//col)\nN_TRAIN = count_data_items(train_files)\nprint(N_TRAIN)\nds = get_training_dataset(train_files)\n\nfor (sample,label) in ds:\n    img = sample['inp1']\n    plt.figure(figsize=(25,int(25*row/col)))\n    for j in range(row*col):\n        plt.subplot(row,col,j+1)\n        plt.title(label[j].numpy())\n        plt.axis('off')\n        plt.imshow(img[j,])\n    plt.show()\n    break\nprint(img.shape)","metadata":{"id":"a3affdd3","outputId":"c0c1fcd6-3d2c-45e8-eadf-f370b9df499f","execution":{"iopub.status.busy":"2022-02-22T14:11:57.118901Z","iopub.execute_input":"2022-02-22T14:11:57.119620Z","iopub.status.idle":"2022-02-22T14:12:22.289931Z","shell.execute_reply.started":"2022-02-22T14:11:57.119582Z","shell.execute_reply":"2022-02-22T14:12:22.287166Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"row = 10; col = 8;\nrow = min(row,config.BATCH_SIZE//col)\nN_TEST = count_data_items(test_files)\nprint(N_TEST)\nds = get_test_dataset(test_files)\n\nfor (img,label) in ds:\n    plt.figure(figsize=(25,int(25*row/col)))\n    for j in range(row*col):\n        plt.subplot(row,col,j+1)\n        plt.title(label[j].numpy())\n        plt.axis('off')\n        plt.imshow(img[j,])\n    plt.show()\n    break\nprint(img.shape)","metadata":{"id":"096877bf","outputId":"a8c3cf72-6438-4afa-ba32-36287f3e0e55","execution":{"iopub.status.busy":"2022-02-22T14:12:22.291389Z","iopub.execute_input":"2022-02-22T14:12:22.291739Z","iopub.status.idle":"2022-02-22T14:12:45.498849Z","shell.execute_reply.started":"2022-02-22T14:12:22.291706Z","shell.execute_reply":"2022-02-22T14:12:45.498030Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Model","metadata":{"id":"85a5ec4c"}},{"cell_type":"code","source":"# Arcmarginproduct class keras layer\nclass ArcMarginProduct(tf.keras.layers.Layer):\n    '''\n    Implements large margin arc distance.\n\n    Reference:\n        https://arxiv.org/pdf/1801.07698.pdf\n        https://github.com/lyakaap/Landmark2019-1st-and-3rd-Place-Solution/\n            blob/master/src/modeling/metric_learning.py\n    '''\n    def __init__(self, n_classes, s=30, m=0.50, easy_margin=False,\n                 ls_eps=0.0, **kwargs):\n\n        super(ArcMarginProduct, self).__init__(**kwargs)\n\n        self.n_classes = n_classes\n        self.s = s\n        self.m = m\n        self.ls_eps = ls_eps\n        self.easy_margin = easy_margin\n        self.cos_m = tf.math.cos(m)\n        self.sin_m = tf.math.sin(m)\n        self.th = tf.math.cos(math.pi - m)\n        self.mm = tf.math.sin(math.pi - m) * m\n\n    def get_config(self):\n\n        config = super().get_config().copy()\n        config.update({\n            'n_classes': self.n_classes,\n            's': self.s,\n            'm': self.m,\n            'ls_eps': self.ls_eps,\n            'easy_margin': self.easy_margin,\n        })\n        return config\n\n    def build(self, input_shape):\n        super(ArcMarginProduct, self).build(input_shape[0])\n\n        self.W = self.add_weight(\n            name='W',\n            shape=(int(input_shape[0][-1]), self.n_classes),\n            initializer='glorot_uniform',\n            dtype='float32',\n            trainable=True,\n            regularizer=None)\n\n    def call(self, inputs):\n        X, y = inputs\n        y = tf.cast(y, dtype=tf.int32)\n        cosine = tf.matmul(\n            tf.math.l2_normalize(X, axis=1),\n            tf.math.l2_normalize(self.W, axis=0)\n        )\n        sine = tf.math.sqrt(1.0 - tf.math.pow(cosine, 2))\n        phi = cosine * self.cos_m - sine * self.sin_m\n        if self.easy_margin:\n            phi = tf.where(cosine > 0, phi, cosine)\n        else:\n            phi = tf.where(cosine > self.th, phi, cosine - self.mm)\n        one_hot = tf.cast(\n            tf.one_hot(y, depth=self.n_classes),\n            dtype=cosine.dtype\n        )\n        if self.ls_eps > 0:\n            one_hot = (1 - self.ls_eps) * one_hot + self.ls_eps / self.n_classes\n\n        output = (one_hot * phi) + ((1.0 - one_hot) * cosine)\n        output *= self.s\n        return output","metadata":{"id":"eed36379","execution":{"iopub.status.busy":"2022-02-22T14:12:45.500173Z","iopub.execute_input":"2022-02-22T14:12:45.500414Z","iopub.status.idle":"2022-02-22T14:12:45.518860Z","shell.execute_reply.started":"2022-02-22T14:12:45.500386Z","shell.execute_reply":"2022-02-22T14:12:45.517923Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"MODEL_NAME = None\nif config.model_type == 'effnetv1':\n    MODEL_NAME = f'effnetv1_b{config.EFF_NET}'\nelif config.model_type == 'effnetv2':\n    MODEL_NAME = f'effnetv2_{config.EFF_NETV2}'\nelse:\n    MODEL_NAME = config.model_type\nconfig.MODEL_NAME = MODEL_NAME\nprint(MODEL_NAME)","metadata":{"id":"65nYxCBRKG1A","outputId":"4f194265-82ca-44ad-a7fc-0dc3bcf9ecbe","execution":{"iopub.status.busy":"2022-02-22T14:12:45.520223Z","iopub.execute_input":"2022-02-22T14:12:45.520446Z","iopub.status.idle":"2022-02-22T14:12:45.537392Z","shell.execute_reply.started":"2022-02-22T14:12:45.520420Z","shell.execute_reply":"2022-02-22T14:12:45.536467Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"EFNS = [efn.EfficientNetB0, efn.EfficientNetB1, efn.EfficientNetB2, efn.EfficientNetB3, \n        efn.EfficientNetB4, efn.EfficientNetB5, efn.EfficientNetB6, efn.EfficientNetB7]\n#from classification_models.keras import Classifiers\n#model, preprocess_input = Classifiers.get('resnet34')\ndef freeze_BN(model):\n    # Unfreeze layers while leaving BatchNorm layers frozen\n    for layer in model.layers:\n        if not isinstance(layer, tf.keras.layers.BatchNormalization):\n            layer.trainable = True\n        else:\n            layer.trainable = False\n\n# Function to create our EfficientNetB3 model\ndef get_model():\n\n    if config.head=='arcface':\n        head = ArcMarginProduct\n    else:\n        assert 1==2, \"INVALID HEAD\"\n    \n    with strategy.scope():\n        \n        margin = head(\n            n_classes = config.N_CLASSES, \n            s = 30, \n            m = 0.3, \n            name=f'head/{config.head}', \n            dtype='float32'\n            )\n\n        #inp = tf.keras.layers.Input(shape = [config.IMAGE_SIZE, config.IMAGE_SIZE, 3], name = 'inp1')\n        label = tf.keras.layers.Input(shape = (), name = 'inp2')\n        #x1 = tf.keras.layers.GlobalAveragePooling2D()(x0.layers[-1].output)  #-3,-7,-9,-15  for EF5    \n        #x2 = tf.keras.layers.GlobalAveragePooling2D()(x0.layers[-3].output) # 803,799,797,791 for EF7\n        #x =  tf.concat([x1,x3,x5],axis = 1)\n        if config.model_type == 'effnetv1':\n            inp = EFNS[config.EFF_NET](weights = 'noisy-student', include_top = False,input_shape = [config.IMAGE_SIZE, config.IMAGE_SIZE, 3])\n            inp.layers[0]._name = 'inp1'\n            #inp.summary()\n            x1=tf.keras.layers.GlobalAveragePooling2D()(inp.layers[-1].output)\n\n            x2=tf.keras.layers.GlobalAveragePooling2D()(inp.layers[-5].output)\n            x3=tf.keras.layers.GlobalAveragePooling2D()(inp.layers[-7].output)\n            x4=tf.keras.layers.GlobalAveragePooling2D()(inp.layers[-13].output)\n            embed =  tf.concat([x1,x2,x3,x4],axis = 1)\n            #embed = tf.keras.layers.GlobalAveragePooling2D()(x)\n        elif config.model_type == 'effnetv2':\n            FEATURE_VECTOR = f'{EFFNETV2_ROOT}/tfhub_models/efficientnetv2-{config.EFF_NETV2}/feature_vector'\n            embed = tfhub.KerasLayer(FEATURE_VECTOR, trainable=True)(inp)\n        else:\n            embed = tfimm.create_model(config.model_type, pretrained=\"timm\")(inp)\n            #embed = tf.keras.layers.GlobalAveragePooling2D()(x)\n\n        embed = tf.keras.layers.Dropout(0.3)(embed)\n        embed = tf.keras.layers.Dense(2048)(embed)\n        #embed = tf.keras.layers.Dropout(0.2)(embed)\n        #embed = tf.keras.layers.Dense(512)(embed)\n        x = margin([embed, label])\n        \n        output = tf.keras.layers.Softmax(dtype='float32')(x)\n        \n        model = tf.keras.models.Model(inputs = [inp.input, label], outputs = [output])\n        embed_model = tf.keras.models.Model(inputs = inp.input, outputs = embed)  \n        \n        opt = tf.keras.optimizers.Adam(learning_rate = config.LR)\n        if config.FREEZE_BATCH_NORM:\n            freeze_BN(model)\n\n        model.compile(\n            optimizer = opt,\n            loss = [tf.keras.losses.SparseCategoricalCrossentropy()],\n            metrics = [tf.keras.metrics.SparseCategoricalAccuracy(),tf.keras.metrics.SparseTopKCategoricalAccuracy(k=5)]\n            ) \n        \n        return model,embed_model","metadata":{"id":"c35c5843","execution":{"iopub.status.busy":"2022-02-22T14:12:45.539323Z","iopub.execute_input":"2022-02-22T14:12:45.539939Z","iopub.status.idle":"2022-02-22T14:12:45.564740Z","shell.execute_reply.started":"2022-02-22T14:12:45.539894Z","shell.execute_reply":"2022-02-22T14:12:45.564054Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def get_lr_callback(plot=False):\n    lr_start   = 0.000001\n    lr_max     = 0.000005 * config.BATCH_SIZE  \n    lr_min     = 0.000001\n    lr_ramp_ep = 4\n    lr_sus_ep  = 0\n    lr_decay   = 0.9\n   \n    def lrfn(epoch):\n        if config.RESUME:\n            epoch = epoch + config.RESUME_EPOCH\n        if epoch < lr_ramp_ep:\n            lr = (lr_max - lr_start) / lr_ramp_ep * epoch + lr_start\n            \n        elif epoch < lr_ramp_ep + lr_sus_ep:\n            lr = lr_max\n            \n        else:\n            lr = (lr_max - lr_min) * lr_decay**(epoch - lr_ramp_ep - lr_sus_ep) + lr_min\n            \n        return lr\n        \n    if plot:\n        epochs = list(range(config.EPOCHS))\n        learning_rates = [lrfn(x) for x in epochs]\n        plt.scatter(epochs,learning_rates)\n        plt.show()\n\n    lr_callback = tf.keras.callbacks.LearningRateScheduler(lrfn, verbose=False)\n    return lr_callback\n\nget_lr_callback(plot=True)","metadata":{"id":"127b8fa0","outputId":"1a693d25-f201-4bc0-bd7e-efd08fce0a6c","execution":{"iopub.status.busy":"2022-02-22T14:12:45.566450Z","iopub.execute_input":"2022-02-22T14:12:45.567028Z","iopub.status.idle":"2022-02-22T14:12:45.791940Z","shell.execute_reply.started":"2022-02-22T14:12:45.566963Z","shell.execute_reply":"2022-02-22T14:12:45.791036Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class Snapshot(tf.keras.callbacks.Callback):\n    \n    def __init__(self,fold,snapshot_epochs=[]):\n        super(Snapshot, self).__init__()\n        self.snapshot_epochs = snapshot_epochs\n        self.fold = fold\n        \n        \n    def on_epoch_end(self, epoch, logs=None):\n        # logs is a dictionary\n#         print(f\"epoch: {epoch}, train_acc: {logs['acc']}, valid_acc: {logs['val_acc']}\")\n        if epoch in self.snapshot_epochs: # your custom condition         \n            self.model.save_weights(config.save_dir+f\"/EF{config.MODEL_NAME}_epoch{epoch}.h5\")\n        self.model.save_weights(config.save_dir+f\"/{config.MODEL_NAME}_last.h5\")","metadata":{"id":"4f632785","execution":{"iopub.status.busy":"2022-02-22T14:12:45.793344Z","iopub.execute_input":"2022-02-22T14:12:45.793573Z","iopub.status.idle":"2022-02-22T14:12:45.800393Z","shell.execute_reply.started":"2022-02-22T14:12:45.793547Z","shell.execute_reply":"2022-02-22T14:12:45.799419Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Train","metadata":{"id":"ffa4b7b9"}},{"cell_type":"code","source":"TRAINING_FILENAMES = [x for i,x in enumerate(train_files) if i%config.FOLDS!=config.FOLD_TO_RUN]\nVALIDATION_FILENAMES = [x for i,x in enumerate(train_files) if i%config.FOLDS==config.FOLD_TO_RUN]\nprint(len(TRAINING_FILENAMES),len(VALIDATION_FILENAMES),count_data_items(TRAINING_FILENAMES),count_data_items(VALIDATION_FILENAMES))","metadata":{"id":"f225e2b0","outputId":"a9573075-9361-449a-8470-ae85e8965a66","execution":{"iopub.status.busy":"2022-02-22T14:12:45.801490Z","iopub.execute_input":"2022-02-22T14:12:45.801831Z","iopub.status.idle":"2022-02-22T14:12:45.819159Z","shell.execute_reply.started":"2022-02-22T14:12:45.801803Z","shell.execute_reply":"2022-02-22T14:12:45.818226Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"if config.DEBUG:\n    TRAINING_FILENAMES = [TRAINING_FILENAMES[0]]\n    VALIDATION_FILENAMES = [VALIDATION_FILENAMES[0]]\n    print(len(TRAINING_FILENAMES),len(VALIDATION_FILENAMES),count_data_items(TRAINING_FILENAMES),count_data_items(VALIDATION_FILENAMES))\n    test_files = [test_files[0]]","metadata":{"id":"90c4bead","execution":{"iopub.status.busy":"2022-02-22T14:12:45.820313Z","iopub.execute_input":"2022-02-22T14:12:45.821079Z","iopub.status.idle":"2022-02-22T14:12:45.831388Z","shell.execute_reply.started":"2022-02-22T14:12:45.821040Z","shell.execute_reply":"2022-02-22T14:12:45.830392Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"seed_everything(config.SEED)\nVERBOSE = 1\ntrain_dataset = get_training_dataset(TRAINING_FILENAMES)\nval_dataset = get_val_dataset(VALIDATION_FILENAMES)\nSTEPS_PER_EPOCH = count_data_items(TRAINING_FILENAMES) // config.BATCH_SIZE\ntrain_logger = tf.keras.callbacks.CSVLogger(config.save_dir+'/training-log-fold-%i.h5.csv'%config.FOLD_TO_RUN)\n# SAVE BEST MODEL EACH FOLD        \nsv_loss = tf.keras.callbacks.ModelCheckpoint(\n    config.save_dir+f\"/{config.MODEL_NAME}_loss_{config.FOLD_TO_RUN}.h5\", monitor='val_loss', verbose=0, save_best_only=True,\n    save_weights_only=True, mode='min', save_freq='epoch')\n# BUILD MODEL\nK.clear_session()\nmodel,embed_model = get_model()\nsnap = Snapshot(fold=config.FOLD_TO_RUN,snapshot_epochs=[5,8])\n#model.summary()\n\nif config.RESUME:   \n    model.load_weights('../input/happywhale-effnet-b7-fork-with-detic-crop-3b3093/effnetv1_b7_loss_4.h5')","metadata":{"id":"3480fcd6","outputId":"2ff1da7d-49e4-468b-d500-02a5571fc9fa","execution":{"iopub.status.busy":"2022-02-22T14:12:45.832926Z","iopub.execute_input":"2022-02-22T14:12:45.833681Z","iopub.status.idle":"2022-02-22T14:14:01.612129Z","shell.execute_reply.started":"2022-02-22T14:12:45.833641Z","shell.execute_reply":"2022-02-22T14:14:01.611008Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import gc\ngc.collect()","metadata":{"id":"unV7Jm-8WRuW","outputId":"2d8ba3b9-912e-45e7-c5c1-fe4ea3ebbfb7","execution":{"iopub.status.busy":"2022-02-22T14:14:01.614805Z","iopub.execute_input":"2022-02-22T14:14:01.615210Z","iopub.status.idle":"2022-02-22T14:14:01.949342Z","shell.execute_reply.started":"2022-02-22T14:14:01.615175Z","shell.execute_reply":"2022-02-22T14:14:01.948277Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print('#### Image Size %i with EfficientNet B%i and batch_size %i'%\n      (config.IMAGE_SIZE,config.EFF_NET,config.BATCH_SIZE))\n\nhistory = model.fit(train_dataset,\n                validation_data = val_dataset,\n                steps_per_epoch = STEPS_PER_EPOCH,\n                epochs = config.EPOCHS,\n                callbacks = [snap,get_lr_callback(),train_logger,sv_loss], \n                verbose = VERBOSE)","metadata":{"id":"5a532288","outputId":"a7758582-8561-4b60-ab83-f5f03d9df621","execution":{"iopub.status.busy":"2022-02-22T14:14:01.951279Z","iopub.execute_input":"2022-02-22T14:14:01.951587Z","iopub.status.idle":"2022-02-22T20:33:04.997481Z","shell.execute_reply.started":"2022-02-22T14:14:01.951556Z","shell.execute_reply":"2022-02-22T20:33:04.994468Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"model.load_weights(config.save_dir+f\"/{config.MODEL_NAME}_loss_{config.FOLD_TO_RUN}.h5\")","metadata":{"id":"7d633c7f","execution":{"iopub.status.busy":"2022-02-22T20:33:05.003671Z","iopub.execute_input":"2022-02-22T20:33:05.005918Z","iopub.status.idle":"2022-02-22T20:33:24.729318Z","shell.execute_reply.started":"2022-02-22T20:33:05.005806Z","shell.execute_reply":"2022-02-22T20:33:24.728273Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Evaluation","metadata":{"id":"ea44925e"}},{"cell_type":"code","source":"def get_ids(filename):\n    ds = get_test_dataset([filename],get_names=True).map(lambda image, image_name: image_name).unbatch()\n    NUM_IMAGES = count_data_items([filename])\n    ids = next(iter(ds.batch(NUM_IMAGES))).numpy().astype('U')\n    return ids\n\ndef get_targets(filename):\n    ds = get_eval_dataset([filename],get_targets=True).map(lambda image, target: target).unbatch()\n    NUM_IMAGES = count_data_items([filename])\n    ids = next(iter(ds.batch(NUM_IMAGES))).numpy()\n    return ids\n\ndef get_embeddings(filename):\n    ds = get_test_dataset([filename],get_names=False)\n    embeddings = embed_model.predict(ds,verbose=0)\n    return embeddings\n\ndef get_predictions(test_df,threshold=0.2):\n    predictions = {}\n    for i,row in tqdm(test_df.iterrows()):\n        if row.image in predictions:\n            if len(predictions[row.image])==5:\n                continue\n            predictions[row.image].append(row.target)\n        elif row.confidence>threshold:\n            predictions[row.image] = [row.target,'new_individual']\n        else:\n            predictions[row.image] = ['new_individual',row.target]\n\n    for x in tqdm(predictions):\n        if len(predictions[x])<5:\n            remaining = [y for y in sample_list if y not in predictions]\n            predictions[x] = predictions[x]+remaining\n            predictions[x] = predictions[x][:5]\n        \n    return predictions\n\ndef map_per_image(label, predictions):\n    \"\"\"Computes the precision score of one image.\n\n    Parameters\n    ----------\n    label : string\n            The true label of the image\n    predictions : list\n            A list of predicted elements (order does matter, 5 predictions allowed per image)\n\n    Returns\n    -------\n    score : double\n    \"\"\"    \n    try:\n        return 1 / (predictions[:5].index(label) + 1)\n    except ValueError:\n        return 0.0\n    \nf = open (\"../input/happywhale-splits/individual_ids.json\", \"r\")\ntarget_encodings = json.loads(f.read())\ntarget_encodings = {target_encodings[x]:x for x in target_encodings}\nsample_list = ['938b7e931166', '5bf17305f073', '7593d2aee842', '7362d7a01d00','956562ff2888']","metadata":{"id":"d836a44d","execution":{"iopub.status.busy":"2022-02-22T20:34:34.243306Z","iopub.execute_input":"2022-02-22T20:34:34.244774Z","iopub.status.idle":"2022-02-22T20:34:34.300961Z","shell.execute_reply.started":"2022-02-22T20:34:34.244688Z","shell.execute_reply":"2022-02-22T20:34:34.300157Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_targets = []\ntrain_embeddings = []\ntrain_data_list=[] \n\nfor filename in tqdm(TRAINING_FILENAMES):\n    embeddings = get_embeddings(filename)\n    with open(config.save_dir+f'train_{filename.split(\"/\")[-1]}_{config.FOLD_TO_RUN}.npy', 'wb') as f:\n      np.save(f, embeddings)    \n    targets = get_targets(filename)\n    train_embeddings.append(embeddings)\n    train_targets.append(targets)\n    train_data_list.append([filename,embeddings])\ntrain_embeddings_df = pd.DataFrame(train_data_list, columns=['filename', 'embeddings'])\ntrain_embeddings_df['FOLD_TO_RUN']=config.FOLD_TO_RUN\ntrain_embeddings_df.to_csv(config.save_dir+f\"/train_embeddings_{config.FOLD_TO_RUN}.csv\",index=False)\ntrain_embeddings = np.concatenate(train_embeddings)\ntrain_targets = np.concatenate(train_targets)","metadata":{"id":"0b7bd225","execution":{"iopub.status.busy":"2022-02-22T20:34:46.108692Z","iopub.execute_input":"2022-02-22T20:34:46.109552Z","iopub.status.idle":"2022-02-22T20:40:36.644249Z","shell.execute_reply.started":"2022-02-22T20:34:46.109502Z","shell.execute_reply":"2022-02-22T20:40:36.643302Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from sklearn.neighbors import NearestNeighbors\nneigh = NearestNeighbors(n_neighbors=config.KNN,metric='cosine')\nneigh.fit(train_embeddings)","metadata":{"id":"4b10b0a6","execution":{"iopub.status.busy":"2022-02-22T20:40:36.646518Z","iopub.execute_input":"2022-02-22T20:40:36.646875Z","iopub.status.idle":"2022-02-22T20:40:36.937982Z","shell.execute_reply.started":"2022-02-22T20:40:36.646825Z","shell.execute_reply":"2022-02-22T20:40:36.936807Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test_ids = []\ntest_nn_distances = []\ntest_nn_idxs = []\nval_targets = []\nval_embeddings = []\nval_data_list=[] \nfor filename in tqdm(VALIDATION_FILENAMES):\n    embeddings = get_embeddings(filename)\n    with open(config.save_dir+f'val_{filename.split(\"/\")[-1]}_{config.FOLD_TO_RUN}.npy', 'wb') as f:\n      np.save(f, embeddings) \n    val_data_list.append([filename,embeddings])\n    targets = get_targets(filename)\n    ids = get_ids(filename)\n    distances,idxs = neigh.kneighbors(embeddings, config.KNN, return_distance=True)\n    test_ids.append(ids)\n    test_nn_idxs.append(idxs)\n    test_nn_distances.append(distances)\n    val_embeddings.append(embeddings)\n    val_targets.append(targets)\nval_embeddings_df = pd.DataFrame(val_data_list, columns=['filename', 'embeddings'])\nval_embeddings_df['FOLD_TO_RUN']=config.FOLD_TO_RUN\nval_embeddings_df.to_csv(config.save_dir+f\"/val_embeddings_{config.FOLD_TO_RUN}.csv\",index=False)\n\ntest_nn_distances = np.concatenate(test_nn_distances)\ntest_nn_idxs = np.concatenate(test_nn_idxs)\ntest_ids = np.concatenate(test_ids)\nval_embeddings = np.concatenate(val_embeddings)\nval_targets = np.concatenate(val_targets)","metadata":{"id":"f11b4473","execution":{"iopub.status.busy":"2022-02-22T20:40:36.939734Z","iopub.execute_input":"2022-02-22T20:40:36.940041Z","iopub.status.idle":"2022-02-22T20:42:54.181723Z","shell.execute_reply.started":"2022-02-22T20:40:36.940001Z","shell.execute_reply":"2022-02-22T20:42:54.180835Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"allowed_targets = set([target_encodings[x] for x in np.unique(train_targets)])\nval_targets_df = pd.DataFrame(np.stack([test_ids,val_targets],axis=1),columns=['image','target'])\nval_targets_df['target'] = val_targets_df['target'].astype(int).map(target_encodings)\nval_targets_df.loc[~val_targets_df.target.isin(allowed_targets),'target'] = 'new_individual'\nval_targets_df.target.value_counts()","metadata":{"id":"d1e251a8","execution":{"iopub.status.busy":"2022-02-22T20:42:54.183732Z","iopub.execute_input":"2022-02-22T20:42:54.184175Z","iopub.status.idle":"2022-02-22T20:42:54.281639Z","shell.execute_reply.started":"2022-02-22T20:42:54.184141Z","shell.execute_reply":"2022-02-22T20:42:54.281067Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test_df = []\nfor i in tqdm(range(len(test_ids))):\n    id_ = test_ids[i]\n    targets = train_targets[test_nn_idxs[i]]\n    distances = test_nn_distances[i]\n    subset_preds = pd.DataFrame(np.stack([targets,distances],axis=1),columns=['target','distances'])\n    subset_preds['image'] = id_\n    test_df.append(subset_preds)\ntest_df = pd.concat(test_df).reset_index(drop=True)\ntest_df['confidence'] = 1-test_df['distances']\ntest_df = test_df.groupby(['image','target']).confidence.max().reset_index()\ntest_df = test_df.sort_values('confidence',ascending=False).reset_index(drop=True)\ntest_df['target'] = test_df['target'].map(target_encodings)\ntest_df.to_csv('val_neighbors.csv')\ntest_df.image.value_counts().value_counts()","metadata":{"id":"9eb82eb7","execution":{"iopub.status.busy":"2022-02-22T20:42:54.282531Z","iopub.execute_input":"2022-02-22T20:42:54.283088Z","iopub.status.idle":"2022-02-22T20:43:06.870605Z","shell.execute_reply.started":"2022-02-22T20:42:54.283059Z","shell.execute_reply":"2022-02-22T20:43:06.869683Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"## Compute CV\nbest_th = 0\nbest_cv = 0\nfor th in [0.1*x for x in range(11)]:\n    all_preds = get_predictions(test_df,threshold=th)\n    cv = 0\n    for i,row in val_targets_df.iterrows():\n        target = row.target\n        preds = all_preds[row.image]\n        val_targets_df.loc[i,th] = map_per_image(target,preds)\n    cv = val_targets_df[th].mean()\n    print(f\"CV at threshold {th}: {cv}\")\n    if cv>best_cv:\n        best_th = th\n        best_cv = cv","metadata":{"id":"85d6cab2","execution":{"iopub.status.busy":"2022-02-22T20:43:06.871773Z","iopub.execute_input":"2022-02-22T20:43:06.872279Z","iopub.status.idle":"2022-02-22T20:48:58.097848Z","shell.execute_reply.started":"2022-02-22T20:43:06.872248Z","shell.execute_reply":"2022-02-22T20:48:58.096821Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(\"Best threshold\",best_th)\nprint(\"Best cv\",best_cv)\nval_targets_df.describe()","metadata":{"id":"e46ccd07","execution":{"iopub.status.busy":"2022-02-22T20:33:24.798448Z","iopub.status.idle":"2022-02-22T20:33:24.798768Z","shell.execute_reply.started":"2022-02-22T20:33:24.798606Z","shell.execute_reply":"2022-02-22T20:33:24.798621Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"## Adjustment: Since Public lb has nearly 10% 'new_individual' (Be Careful for private LB)\nval_targets_df['is_new_individual'] = val_targets_df.target=='new_individual'\nprint(val_targets_df.is_new_individual.value_counts().to_dict())\nval_scores = val_targets_df.groupby('is_new_individual').mean().T\nval_scores['adjusted_cv'] = val_scores[True]*0.1+val_scores[False]*0.9\nbest_threshold_adjusted = val_scores['adjusted_cv'].idxmax()\nprint(\"best_threshold\",best_threshold_adjusted)\nval_scores","metadata":{"id":"3292cf2e","execution":{"iopub.status.busy":"2022-02-22T20:48:58.099391Z","iopub.execute_input":"2022-02-22T20:48:58.100326Z","iopub.status.idle":"2022-02-22T20:48:58.138510Z","shell.execute_reply.started":"2022-02-22T20:48:58.100285Z","shell.execute_reply":"2022-02-22T20:48:58.137570Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Inference","metadata":{"id":"68ec2024"}},{"cell_type":"code","source":"train_embeddings = np.concatenate([train_embeddings,val_embeddings])\ntrain_targets = np.concatenate([train_targets,val_targets])\nprint(train_embeddings.shape,train_targets.shape)","metadata":{"id":"0bf74cb1","execution":{"iopub.status.busy":"2022-02-22T20:48:58.139890Z","iopub.execute_input":"2022-02-22T20:48:58.140644Z","iopub.status.idle":"2022-02-22T20:48:58.577748Z","shell.execute_reply.started":"2022-02-22T20:48:58.140600Z","shell.execute_reply":"2022-02-22T20:48:58.576574Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from sklearn.neighbors import NearestNeighbors\nneigh = NearestNeighbors(n_neighbors=config.KNN,metric='cosine')\nneigh.fit(train_embeddings)","metadata":{"id":"47ab017f","execution":{"iopub.status.busy":"2022-02-22T20:48:58.579140Z","iopub.execute_input":"2022-02-22T20:48:58.579445Z","iopub.status.idle":"2022-02-22T20:48:58.676566Z","shell.execute_reply.started":"2022-02-22T20:48:58.579403Z","shell.execute_reply":"2022-02-22T20:48:58.675724Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test_ids = []\ntest_nn_distances = []\ntest_nn_idxs = []\ntest_data_list=[] \nfor filename in tqdm(test_files):\n    embeddings = get_embeddings(filename)\n    with open(config.save_dir+f'test_{filename.split(\"/\")[-1]}_{config.FOLD_TO_RUN}.npy', 'wb') as f:\n      np.save(f, embeddings) \n    test_data_list.append([filename,embeddings])\n    ids = get_ids(filename)\n    distances,idxs = neigh.kneighbors(embeddings, config.KNN, return_distance=True)\n    test_ids.append(ids)\n    test_nn_idxs.append(idxs)\n    test_nn_distances.append(distances)\n\ntest_embeddings_df = pd.DataFrame(test_data_list, columns=['filename', 'embeddings'])\ntest_embeddings_df['FOLD_TO_RUN']=config.FOLD_TO_RUN\ntest_embeddings_df.to_csv(config.save_dir+f\"/test_embeddings_{config.FOLD_TO_RUN}.csv\",index=False)\n\n\n\ntest_nn_distances = np.concatenate(test_nn_distances)\ntest_nn_idxs = np.concatenate(test_nn_idxs)\ntest_ids = np.concatenate(test_ids)","metadata":{"id":"81d5332c","execution":{"iopub.status.busy":"2022-02-22T20:48:58.679103Z","iopub.execute_input":"2022-02-22T20:48:58.679422Z","iopub.status.idle":"2022-02-22T20:57:40.325588Z","shell.execute_reply.started":"2022-02-22T20:48:58.679391Z","shell.execute_reply":"2022-02-22T20:57:40.324687Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"sample_submission = pd.read_csv(\"../input/happy-whale-and-dolphin/sample_submission.csv\",index_col='image')\nprint(len(test_ids),len(sample_submission))\ntest_df = []\nfor i in tqdm(range(len(test_ids))):\n    id_ = test_ids[i]\n    targets = train_targets[test_nn_idxs[i]]\n    distances = test_nn_distances[i]\n    subset_preds = pd.DataFrame(np.stack([targets,distances],axis=1),columns=['target','distances'])\n    subset_preds['image'] = id_\n    test_df.append(subset_preds)\ntest_df = pd.concat(test_df).reset_index(drop=True)\ntest_df['confidence'] = 1-test_df['distances']\ntest_df = test_df.groupby(['image','target']).confidence.max().reset_index()\ntest_df = test_df.sort_values('confidence',ascending=False).reset_index(drop=True)\ntest_df['target'] = test_df['target'].map(target_encodings)\ntest_df.to_csv('test_neighbors.csv')\ntest_df.image.value_counts().value_counts()","metadata":{"id":"f0c3207e","execution":{"iopub.status.busy":"2022-02-22T21:00:19.989767Z","iopub.execute_input":"2022-02-22T21:00:19.990831Z","iopub.status.idle":"2022-02-22T21:01:01.795175Z","shell.execute_reply.started":"2022-02-22T21:00:19.990784Z","shell.execute_reply":"2022-02-22T21:01:01.794156Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"sample_list = ['938b7e931166', '5bf17305f073', '7593d2aee842', '7362d7a01d00','956562ff2888']","metadata":{"id":"16e43dbd","execution":{"iopub.status.busy":"2022-02-22T21:01:01.797362Z","iopub.execute_input":"2022-02-22T21:01:01.798149Z","iopub.status.idle":"2022-02-22T21:01:01.804609Z","shell.execute_reply.started":"2022-02-22T21:01:01.798108Z","shell.execute_reply":"2022-02-22T21:01:01.803424Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"predictions = {}\nfor i,row in tqdm(test_df.iterrows()):\n    if row.image in predictions:\n        if len(predictions[row.image])==5:\n            continue\n        predictions[row.image].append(row.target)\n    elif row.confidence>best_threshold_adjusted:\n        predictions[row.image] = [row.target,'new_individual']\n    else:\n        predictions[row.image] = ['new_individual',row.target]\n        \nfor x in tqdm(predictions):\n    if len(predictions[x])<5:\n        remaining = [y for y in sample_list if y not in predictions]\n        predictions[x] = predictions[x]+remaining\n        predictions[x] = predictions[x][:5]\n    predictions[x] = ' '.join(predictions[x])\n    \npredictions = pd.Series(predictions).reset_index()\npredictions.columns = ['image','predictions']\npredictions.to_csv(config.save_dir+f\"/submission.csv\",index=False)\npredictions.head()","metadata":{"id":"7194d653","execution":{"iopub.status.busy":"2022-02-22T21:01:01.805937Z","iopub.execute_input":"2022-02-22T21:01:01.806199Z","iopub.status.idle":"2022-02-22T21:02:25.194051Z","shell.execute_reply.started":"2022-02-22T21:01:01.806172Z","shell.execute_reply":"2022-02-22T21:02:25.193087Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{"id":"dTp_AcY5GGaC"},"execution_count":null,"outputs":[]}]}