{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.13","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"tpu1vmV38","dataSources":[{"sourceId":21154,"databundleVersionId":1243559,"sourceType":"competition"},{"sourceId":1138814,"sourceType":"datasetVersion","datasetId":601927}],"dockerImageVersionId":30132,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import sys\n!git clone https://github.com/rishigami/Swin-Transformer-TF\nsys.path.append('/kaggle/working/Swin-Transformer-TF')","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import math, re, os, random, gc\nimport numpy as np\nimport pandas as pd\nfrom matplotlib import pyplot as plt\nfrom sklearn.metrics import f1_score, precision_score, recall_score, confusion_matrix\nimport tensorflow as tf\nfrom tensorflow_addons.metrics import F1Score \nfrom tensorflow.keras import layers as L\nfrom tensorflow.keras import backend as K\nfrom tensorflow.keras import callbacks\nfrom tensorflow.keras import applications as tf_applications\nfrom tensorflow.keras.optimizers import Adam, Nadam, Adamax\nfrom tensorflow.keras.layers import Dense, Dropout, Conv2D, LayerNormalization, GlobalAveragePooling1D\nfrom kaggle_datasets import KaggleDatasets\nfrom swintransformer import SwinTransformer\nimport warnings\nimport sys\nimport ctypes","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def clean_memory():\n    libc = ctypes.CDLL(\"libc.so.6\")\n    libc.malloc_trim(0)\n    gc.collect()\n    \ndef seed_everything(seed=36):\n    random.seed(seed)\n    np.random.seed(seed)\n    os.environ['PYTHONHASHSEED'] = str(seed)\n    tf.random.set_seed(seed)\n\ndef accelerator():\n    tf.config.optimizer.set_jit(True)\n    print(tf.config.optimizer.get_jit())\n    \n    tf.config.optimizer.set_experimental_options({'disable_model_pruning': True,'scoped_allocator_optimization': True,'implementation_selector': True, 'auto_parallel':True,'constant_folding': True, 'shape_optimization':True, 'remapping':True,'arithmetic_optimization': True, 'dependency_optimization':True, 'function_optimization':True, 'loop_optimization':True})\n    print(tf.config.optimizer.get_experimental_options())\n\ndef get_strategy():\n    try:\n        tpu = tf.distribute.cluster_resolver.TPUClusterResolver()\n        tf.config.experimental_connect_to_cluster(tpu)\n        tf.tpu.experimental.initialize_tpu_system(tpu)\n        strategy = tf.distribute.experimental.TPUStrategy(tpu)\n    except ValueError:\n        strategy = tf.distribute.get_strategy()\n    return strategy\n\naccelerator()\nclean_memory()\nseed_everything(36)\nwarnings.filterwarnings('ignore')","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"strategy = get_strategy()\nAUTO = tf.data.AUTOTUNE\nAUTOTUNE = tf.data.AUTOTUNE","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"IMAGE_SIZE = [224, 224]\nEPOCHS = 36\nBATCH_SIZE = 16 * strategy.num_replicas_in_sync\nSWIN_TYPE = 'tiny'\n","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"GCS_DS_PATH = KaggleDatasets().get_gcs_path('tpu-getting-started')\nGCS_DS_PATH_EXT = KaggleDatasets().get_gcs_path('tf-flower-photo-tfrec')\nGCS_PATH_SELECT = { # available image sizes\n    192: GCS_DS_PATH + '/tfrecords-jpeg-192x192',\n    224: GCS_DS_PATH + '/tfrecords-jpeg-224x224',\n    331: GCS_DS_PATH + '/tfrecords-jpeg-331x331',\n    512: GCS_DS_PATH + '/tfrecords-jpeg-512x512'\n}\nGCS_PATH = GCS_PATH_SELECT[IMAGE_SIZE[0]]\nGCS_PATH_SELECT_EXT = {\n    192: '/tfrecords-jpeg-192x192',\n    224: '/tfrecords-jpeg-224x224',\n    331: '/tfrecords-jpeg-331x331',\n    512: '/tfrecords-jpeg-512x512'\n}\nGCS_PATH_EXT = GCS_PATH_SELECT_EXT[IMAGE_SIZE[0]]\nIMAGENET_FILES = tf.io.gfile.glob(GCS_DS_PATH_EXT + '/imagenet' + GCS_PATH_EXT + '/*.tfrec')\nINATURELIST_FILES = tf.io.gfile.glob(GCS_DS_PATH_EXT + '/inaturalist' + GCS_PATH_EXT + '/*.tfrec')\nOPENIMAGE_FILES = tf.io.gfile.glob(GCS_DS_PATH_EXT + '/openimage' + GCS_PATH_EXT + '/*.tfrec')\nOXFORD_FILES = tf.io.gfile.glob(GCS_DS_PATH_EXT + '/oxford_102' + GCS_PATH_EXT + '/*.tfrec')\nTENSORFLOW_FILES = tf.io.gfile.glob(GCS_DS_PATH_EXT + '/tf_flowers' + GCS_PATH_EXT + '/*.tfrec')\nADDITIONAL_TRAINING_FILENAMES = IMAGENET_FILES + INATURELIST_FILES + OPENIMAGE_FILES + OXFORD_FILES + TENSORFLOW_FILES  \nCLASSES = ['pink primrose',    'hard-leaved pocket orchid', 'canterbury bells', 'sweet pea',     'wild geranium',     'tiger lily',           'moon orchid',              'bird of paradise', 'monkshood',        'globe thistle',         # 00 - 09\n           'snapdragon',       \"colt's foot\",               'king protea',      'spear thistle', 'yellow iris',       'globe-flower',         'purple coneflower',        'peruvian lily',    'balloon flower',   'giant white arum lily', # 10 - 19\n           'fire lily',        'pincushion flower',         'fritillary',       'red ginger',    'grape hyacinth',    'corn poppy',           'prince of wales feathers', 'stemless gentian', 'artichoke',        'sweet william',         # 20 - 29\n           'carnation',        'garden phlox',              'love in the mist', 'cosmos',        'alpine sea holly',  'ruby-lipped cattleya', 'cape flower',              'great masterwort', 'siam tulip',       'lenten rose',           # 30 - 39\n           'barberton daisy',  'daffodil',                  'sword lily',       'poinsettia',    'bolero deep blue',  'wallflower',           'marigold',                 'buttercup',        'daisy',            'common dandelion',      # 40 - 49\n           'petunia',          'wild pansy',                'primula',          'sunflower',     'lilac hibiscus',    'bishop of llandaff',   'gaura',                    'geranium',         'orange dahlia',    'pink-yellow dahlia',    # 50 - 59\n           'cautleya spicata', 'japanese anemone',          'black-eyed susan', 'silverbush',    'californian poppy', 'osteospermum',         'spring crocus',            'iris',             'windflower',       'tree poppy',            # 60 - 69\n           'gazania',          'azalea',                    'water lily',       'rose',          'thorn apple',       'morning glory',        'passion flower',           'lotus',            'toad lily',        'anthurium',             # 70 - 79\n           'frangipani',       'clematis',                  'hibiscus',         'columbine',     'desert-rose',       'tree mallow',          'magnolia',                 'cyclamen ',        'watercress',       'canna lily',            # 80 - 89\n           'hippeastrum ',     'bee balm',                  'pink quill',       'foxglove',      'bougainvillea',     'camellia',             'mallow',                   'mexican petunia',  'bromelia',         'blanket flower',        # 90 - 99\n           'trumpet creeper',  'blackberry lily',           'common tulip',     'wild rose']                          # 100 - 102\n\nTRAINING_FILENAMES = tf.io.gfile.glob(GCS_PATH + '/train/*.tfrec')\nVALIDATION_FILENAMES = tf.io.gfile.glob(GCS_PATH + '/val/*.tfrec')\nTEST_FILENAMES = tf.io.gfile.glob(GCS_PATH + '/test/*.tfrec') # predictions on this dataset should be submitted for the competition \n\nTRAINING_FILENAMES = TRAINING_FILENAMES + ADDITIONAL_TRAINING_FILENAMES","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# numpy and matplotlib defaults\nnp.set_printoptions(threshold=15, linewidth=80)\nwith strategy.scope():\n    def batch_to_numpy_images_and_labels(data):\n        images, labels = data\n        numpy_images = images.numpy()\n        numpy_labels = labels.numpy()\n        if numpy_labels.dtype == object: # binary string in this case, these are image ID strings\n            numpy_labels = [None for _ in enumerate(numpy_images)]\n        # If no labels, only image IDs, return None for labels (this is the case for test data)\n        return numpy_images, numpy_labels\n\n    def title_from_label_and_target(label, correct_label):\n        if correct_label is None:\n            return CLASSES[label], True\n        correct = (label == correct_label)\n        return \"{} [{}{}{}]\".format(CLASSES[label], 'OK' if correct else 'NO', u\"\\u2192\" if not correct else '',\n                                    CLASSES[correct_label] if not correct else ''), correct\n\n    def display_one_flower(image, title, subplot, red=False, titlesize=16):\n        plt.subplot(*subplot)\n        plt.axis('off')\n        plt.imshow(image)\n        if len(title) > 0:\n            plt.title(title, fontsize=int(titlesize) if not red else int(titlesize/1.2), color='red' if red else 'black', fontdict={'verticalalignment':'center'}, pad=int(titlesize/1.5))\n        return (subplot[0], subplot[1], subplot[2]+1)\n\n    def display_batch_of_images(databatch, predictions=None):\n        # data\n        images, labels = batch_to_numpy_images_and_labels(databatch)\n        if labels is None:\n            labels = [None for _ in enumerate(images)]\n        # auto-squaring: this will drop data that does not fit into square or square-ish rectangle\n        rows = int(math.sqrt(len(images)))\n        cols = len(images)//rows\n        # size and spacing\n        FIGSIZE = 13.0\n        SPACING = 0.1\n        subplot=(rows,cols,1)\n        if rows < cols:\n            plt.figure(figsize=(FIGSIZE,FIGSIZE/cols*rows))\n        else:\n            plt.figure(figsize=(FIGSIZE/rows*cols,FIGSIZE))\n        # display\n        for i, (image, label) in enumerate(zip(images[:rows*cols], labels[:rows*cols])):\n            title = '' if label is None else CLASSES[label]\n            correct = True\n            if predictions is not None:\n                title, correct = title_from_label_and_target(predictions[i], label)\n            dynamic_titlesize = FIGSIZE*SPACING/max(rows,cols)*40+3 # magic formula tested to work from 1x1 to 10x10 images\n            subplot = display_one_flower(image, title, subplot, not correct, titlesize=dynamic_titlesize)\n        #layout\n        plt.tight_layout()\n        if label is None and predictions is None:\n            plt.subplots_adjust(wspace=0, hspace=0)\n        else:\n            plt.subplots_adjust(wspace=SPACING, hspace=SPACING)\n        plt.show()\n\n    def display_confusion_matrix(cmat, score, precision, recall):\n        plt.figure(figsize=(15,15))\n        ax = plt.gca()\n        ax.matshow(cmat, cmap='Reds')\n        ax.set_xticks(range(len(CLASSES)))\n        ax.set_xticklabels(CLASSES, fontdict={'fontsize': 7})\n        plt.setp(ax.get_xticklabels(), rotation=45, ha=\"left\", rotation_mode=\"anchor\")\n        ax.set_yticks(range(len(CLASSES)))\n        ax.set_yticklabels(CLASSES, fontdict={'fontsize': 7})\n        plt.setp(ax.get_yticklabels(), rotation=45, ha=\"right\", rotation_mode=\"anchor\")\n        titlestring = \"\"\n        if score is not None:\n            titlestring += 'f1 = {:.3f} '.format(score)\n        if precision is not None:\n            titlestring += '\\nprecision = {:.3f} '.format(precision)\n        if recall is not None:\n            titlestring += '\\nrecall = {:.3f} '.format(recall)\n        if len(titlestring) > 0:\n            ax.text(101, 1, titlestring, fontdict={'fontsize': 18, 'horizontalalignment':'right', 'verticalalignment':'top', 'color':'#804040'})\n        plt.show()\n\n    def display_training_curves(training, validation, title, subplot):\n        if subplot%10==1: # set up the subplots on the first call\n            plt.subplots(figsize=(10,10), facecolor='#F0F0F0')\n            plt.tight_layout()\n        ax = plt.subplot(subplot)\n        ax.set_facecolor('#F8F8F8')\n        ax.plot(training)\n        ax.plot(validation)\n        ax.set_title('model '+ title)\n        ax.set_ylabel(title)\n        #ax.set_ylim(0.28,1.05)\n        ax.set_xlabel('epoch')\n        ax.legend(['train', 'valid.'])","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    def random_erasing(img, sl=0.1, sh=0.2, rl=0.4):\n        p=random.random()\n        if p>=0.0 and p<=0.5:\n            w, h, c = IMAGE_SIZE[0], IMAGE_SIZE[1], 3\n            origin_area = tf.cast(h*w, tf.float32)\n            e_size_l = tf.cast(tf.round(tf.sqrt(origin_area * sl * rl)), tf.int32)\n            e_size_h = tf.cast(tf.round(tf.sqrt(origin_area * sh / rl)), tf.int32)\n            e_height_h = tf.minimum(e_size_h, h)\n            e_width_h = tf.minimum(e_size_h, w)\n            erase_height = tf.random.uniform(shape=[], minval=e_size_l, maxval=e_height_h, dtype=tf.int32)\n            erase_width = tf.random.uniform(shape=[], minval=e_size_l, maxval=e_width_h, dtype=tf.int32)\n            erase_area = tf.zeros(shape=[erase_height, erase_width, c])\n            erase_area = tf.cast(erase_area, tf.uint8)\n            pad_h = h - erase_height\n            pad_top = tf.random.uniform(shape=[], minval=0, maxval=pad_h, dtype=tf.int32)\n            pad_bottom = pad_h - pad_top\n            pad_w = w - erase_width\n            pad_left = tf.random.uniform(shape=[], minval=0, maxval=pad_w, dtype=tf.int32)\n            pad_right = pad_w - pad_left\n            erase_mask = tf.pad([erase_area], [[0,0],[pad_top, pad_bottom], [pad_left, pad_right], [0,0]], constant_values=1)\n            erase_mask = tf.squeeze(erase_mask, axis=0)\n            erased_img = tf.multiply(tf.cast(img,tf.float32), tf.cast(erase_mask, tf.float32))\n            return tf.cast(erased_img, img.dtype)\n        else:\n            return tf.cast(img, img.dtype)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    def decode_image(image_data):\n        image = tf.image.decode_jpeg(image_data, channels=3)\n        image = tf.cast(image, tf.float32) / 255.0  # convert image to floats in [0, 1] range\n        image = tf.reshape(image, [*IMAGE_SIZE, 3]) # explicit size needed for TPU\n        return image\n    \n    def transform(image,label):\n        DIM = IMAGE_SIZE[0]\n        XDIM = DIM%2 #fix for size 331\n        rot = 15. * tf.random.normal([1],dtype='float32')\n        shr = 5. * tf.random.normal([1],dtype='float32') \n        h_zoom = 1.0 + tf.random.normal([1],dtype='float32')/10.\n        w_zoom = 1.0 + tf.random.normal([1],dtype='float32')/10.\n        h_shift = 16. * tf.random.normal([1],dtype='float32') \n        w_shift = 16. * tf.random.normal([1],dtype='float32') \n        m = get_mat(rot,shr,h_zoom,w_zoom,h_shift,w_shift) \n        x = tf.repeat( tf.range(DIM//2,-DIM//2,-1), DIM )\n        y = tf.tile( tf.range(-DIM//2,DIM//2),[DIM] )\n        z = tf.ones([DIM*DIM],dtype='int32')\n        idx = tf.stack( [x,y,z] )\n        idx2 = K.dot(m,tf.cast(idx,dtype='float32'))\n        idx2 = K.cast(idx2,dtype='int32')\n        idx2 = K.clip(idx2,-DIM//2+XDIM+1,DIM//2)\n        idx3 = tf.stack( [DIM//2-idx2[0,], DIM//2-1+idx2[1,]] )\n        d = tf.gather_nd(image,tf.transpose(idx3))\n        return tf.reshape(d,[DIM,DIM,3]),label\n    \n    def onehot(image,label):\n        return image,tf.one_hot(label, len(CLASSES))\n\n    def read_labeled_tfrecord(example):\n        LABELED_TFREC_FORMAT = {\n            \"image\": tf.io.FixedLenFeature([], tf.string), # tf.string means bytestring\n            \"class\": tf.io.FixedLenFeature([], tf.int64),  # shape [] means single element\n        }\n        example = tf.io.parse_single_example(example, LABELED_TFREC_FORMAT)\n        image = decode_image(example['image'])\n        label = tf.cast(example['class'], tf.int32)\n        return image, label # returns a dataset of (image, label) pairs\n\n    def read_unlabeled_tfrecord(example):\n        UNLABELED_TFREC_FORMAT = {\n            \"image\": tf.io.FixedLenFeature([], tf.string), # tf.string means bytestring\n            \"id\": tf.io.FixedLenFeature([], tf.string),  # shape [] means single element\n        }\n        example = tf.io.parse_single_example(example, UNLABELED_TFREC_FORMAT)\n        image = decode_image(example['image'])\n        idnum = example['id']\n        return image, idnum # returns a dataset of image(s)\n\n    def load_dataset(filenames, labeled=True, ordered=False):\n        ignore_order = tf.data.Options()\n        if not ordered:\n            ignore_order.experimental_deterministic = False # disable order, increase speed\n\n        dataset = tf.data.TFRecordDataset(filenames, num_parallel_reads=AUTO) # automatically interleaves reads from multiple files\n        dataset = dataset.with_options(ignore_order) # uses data as soon as it streams in, rather than in its original order\n        dataset = dataset.map(read_labeled_tfrecord if labeled else read_unlabeled_tfrecord, num_parallel_calls=AUTO)\n        # returns a dataset of (image, label) pairs if labeled=True or (image, id) pairs if labeled=False\n        return dataset\n\n    def data_augment(image, label):\n        image = tf.image.random_flip_left_right(image)\n        image = tf.image.random_brightness(image, 0.2)\n        image = tf.image.random_contrast(image, 0.8, 1.2)\n        image = tf.image.random_saturation(image, 0.8, 1.2)\n        image = random_erasing(image)\n        return image, label\n\n    def data_hflip(image, idnum):\n        image = tf.image.random_flip_left_right(image)\n        return image, idnum\n\n    def get_training_dataset(do_onehot=False):\n        dataset = load_dataset(TRAINING_FILENAMES, labeled=True)\n        dataset = dataset.map(data_augment, num_parallel_calls=AUTO)\n        if do_onehot:\n            dataset = dataset.map(onehot, num_parallel_calls=AUTO)\n        dataset = dataset.repeat() # the training dataset must repeat for several epochs\n        dataset = dataset.shuffle(2048)\n        dataset = dataset.batch(BATCH_SIZE)\n        dataset = dataset.prefetch(AUTO) # prefetch next batch while training (autotune prefetch buffer size)\n        return dataset\n\n    def get_validation_dataset(ordered=False, do_onehot=False):\n        dataset = load_dataset(VALIDATION_FILENAMES, labeled=True, ordered=ordered)\n        if do_onehot:\n            dataset = dataset.map(onehot, num_parallel_calls=AUTO)\n        dataset = dataset.batch(BATCH_SIZE)\n        dataset = dataset.cache()\n        dataset = dataset.prefetch(AUTO)\n        return dataset\n\n    def get_test_dataset(ordered=False, augmented=False):\n        dataset = load_dataset(TEST_FILENAMES, labeled=False, ordered=ordered)\n        dataset = dataset.map(data_hflip, num_parallel_calls=AUTO)\n        dataset = dataset.batch(BATCH_SIZE)\n        dataset = dataset.prefetch(AUTO)\n        return dataset\n\n    def count_data_items(filenames):\n        n = [int(re.compile(r\"-([0-9]*)\\.\").search(filename).group(1)) for filename in filenames]\n        return np.sum(n)\n\nNUM_TRAINING_IMAGES = count_data_items(TRAINING_FILENAMES)\nNUM_VALIDATION_IMAGES = count_data_items(VALIDATION_FILENAMES)\nNUM_TEST_IMAGES = count_data_items(TEST_FILENAMES)\nSTEPS_PER_EPOCH = NUM_TRAINING_IMAGES // BATCH_SIZE\nVALIDATION_STEPS = -(-NUM_VALIDATION_IMAGES // BATCH_SIZE) # The \"-(-//)\" trick rounds up instead of down :-)\nTEST_STEPS = -(-NUM_TEST_IMAGES // BATCH_SIZE)             # The \"-(-//)\" trick rounds up instead of down :-)\nprint(f'Dataset: {NUM_TRAINING_IMAGES} training images, {NUM_VALIDATION_IMAGES} validation images, {NUM_TEST_IMAGES} unlabeled test images')","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    def get_lr_callback(plot_schedule=False):\n        LR_START = 0.00001\n        LR_MAX = 0.00005 * strategy.num_replicas_in_sync\n        LR_MIN = 0.00001\n        LR_RAMPUP_EPOCHS = 5\n        LR_SUSTAIN_EPOCHS = 0\n        LR_EXP_DECAY = .8\n\n        def lrfn(epoch):\n            if epoch < LR_RAMPUP_EPOCHS:\n                lr = (LR_MAX - LR_START) / LR_RAMPUP_EPOCHS * epoch + LR_START\n            elif epoch < LR_RAMPUP_EPOCHS + LR_SUSTAIN_EPOCHS:\n                lr = LR_MAX\n            else:\n                lr = (LR_MAX - LR_MIN) * LR_EXP_DECAY**(epoch - LR_RAMPUP_EPOCHS - LR_SUSTAIN_EPOCHS) + LR_MIN\n            return lr\n\n        return callbacks.LearningRateScheduler(lrfn, verbose=1)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"CFGS = {\n    'swin_tiny_224': dict(input_size=(224, 224), window_size=7, embed_dim=96, depths=[2, 2, 6, 2], num_heads=[3, 6, 12, 24]),\n    'swin_tiny_331': dict(input_size=(331, 331), window_size=12, embed_dim=128, depths=[2, 2, 18, 2], num_heads=[3, 6, 12, 24]),\n    'swin_base_224': dict(input_size=(224, 224), window_size=7, embed_dim=128, depths=[2, 2, 18, 2], num_heads=[4, 8, 16, 32]),\n    'swin_base_331': dict(input_size=(331, 331), window_size=12, embed_dim=128, depths=[2, 2, 18, 2], num_heads=[4, 8, 16, 32]),\n    'swin_large_224': dict(input_size=(224, 224), window_size=7, embed_dim=192, depths=[2, 2, 18, 2], num_heads=[6, 12, 24, 48]),\n    'swin_large_331': dict(input_size=(331, 331), window_size=12, embed_dim=192, depths=[2, 2, 18, 2], num_heads=[6, 12, 24, 48])\n}\n\ndef load_and_fit_model(print_summary=False):\n    with strategy.scope():\n        model = tf.keras.Sequential([\n            SwinTransformer(f'swin_{SWIN_TYPE}_{IMAGE_SIZE[0]}', include_top=False, pretrained=True, cfgs=CFGS),\n            L.Dense(len(CLASSES), activation='softmax')\n        ])\n\n        model.compile(\n            optimizer=Adam(beta_1=0.5),\n            loss = 'categorical_crossentropy',\n            metrics=[F1Score(len(CLASSES), average='macro')],\n        )\n    os.makedirs('checkpoints', exist_ok=True)\n    lr_callback = get_lr_callback()\n    chk_callback = callbacks.ModelCheckpoint(f'checkpoints/{model_name}_best.h5',save_weights_only=True, monitor='val_f1_score',mode='max', save_best_only=True, verbose=1)\n    ton = tf.keras.callbacks.TerminateOnNaN()\n    _ = model.fit(get_training_dataset(do_onehot=True), \n                  steps_per_epoch=STEPS_PER_EPOCH, \n                  epochs=EPOCHS, \n                  validation_data=get_validation_dataset(do_onehot=True),\n                  validation_steps=VALIDATION_STEPS,\n                  callbacks=[lr_callback, chk_callback, ton],\n                  verbose=1)\n    model.load_weights(f'checkpoints/{model_name}_best.h5')\n    return model\n    ","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    model = load_and_fit_model()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# def find_best_alpha(valid_dataset, model_lst):\n#     images_ds = valid_dataset.map(lambda image, label: image)\n#     labels_ds = valid_dataset.map(lambda image, label: label).unbatch()\n#     y_true = next(iter(labels_ds.batch(NUM_VALIDATION_IMAGES))).numpy() # get everything as one batch\n#     p = []\n#     for model in model_lst:\n#         p.append(model.predict(images_ds))\n#     scores = []\n#     for alpha in np.linspace(0,1,100):\n#         preds = np.argmax(alpha*p[0]+(1-alpha)*p[1], axis=-1)\n#         scores.append(f1_score(y_true, preds, labels=range(len(CLASSES)), average='macro'))\n#     best_alpha = np.argmax(scores)/100\n#     return best_alpha","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# valid_ds = get_validation_dataset(ordered=True) # since we are splitting the dataset and iterating separately on images and labels, order matters.\n# alpha = find_best_alpha(valid_ds, models)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# def predict_ensemble(dataset, model_lst, alpha, steps):\n#     images_ds = dataset.map(lambda image, idnum: image)\n#     probs = []\n#     for model in model_lst:\n#         p = model.predict(images_ds,verbose=0, steps=steps)\n#         probs.append(p)\n#     preds = np.argmax(alpha*probs[0] + (1-alpha)*probs[1], axis=-1)\n#     return preds","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# cm_predictions = predict_ensemble(valid_ds, models, alpha, steps=VALIDATION_STEPS)\n# labels_ds = valid_ds.map(lambda image, label: label).unbatch()\n# cm_correct_labels = next(iter(labels_ds.batch(NUM_VALIDATION_IMAGES))).numpy() # get everything as one batch\n# cmat = confusion_matrix(cm_correct_labels, cm_predictions, labels=range(len(CLASSES)))\n# score = f1_score(cm_correct_labels, cm_predictions, labels=range(len(CLASSES)), average='macro')\n# precision = precision_score(cm_correct_labels, cm_predictions, labels=range(len(CLASSES)), average='macro')\n# recall = recall_score(cm_correct_labels, cm_predictions, labels=range(len(CLASSES)), average='macro')\n# display_confusion_matrix(cmat, score, precision, recall)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# test_ds = get_test_dataset(ordered=True) # since we are splitting the dataset and iterating separately on images and ids, order matters.\n# predictions = predict_ensemble(test_ds, models, alpha, steps=TEST_STEPS)\n# test_ids_ds = test_ds.map(lambda image, idnum: idnum).unbatch()\n# test_ids = next(iter(test_ids_ds.batch(NUM_TEST_IMAGES))).numpy().astype('U') # all in one batch\n# sub_df = pd.DataFrame({'id': test_ids, 'label': predictions})\n# sub_df.to_csv('submission.csv', index=False)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def predict(dataset, model):\n    print('Calculating predictions...')\n    images_ds = dataset.map(lambda image, idnum: image)\n    preds = model.predict(images_ds,verbose=0)\n    preds = np.argmax(preds, axis=1)\n    return preds\n\n# valid_ds = get_validation_dataset(ordered=True)\n# cm_predictions = predict(valid_ds, model)\n\n# labels_ds = valid_ds.map(lambda image, label: label).unbatch()\n# cm_correct_labels = next(iter(labels_ds.batch(NUM_VALIDATION_IMAGES))).numpy() # get everything as one batch\n\n# cmat = confusion_matrix(cm_correct_labels, cm_predictions, labels=range(len(CLASSES)))\n# score = f1_score(cm_correct_labels, cm_predictions, labels=range(len(CLASSES)), average='macro')\n# precision = precision_score(cm_correct_labels, cm_predictions, labels=range(len(CLASSES)), average='macro')\n# recall = recall_score(cm_correct_labels, cm_predictions, labels=range(len(CLASSES)), average='macro')\n# #cmat = (cmat.T / cmat.sum(axis=1)).T # normalized\n# display_confusion_matrix(cmat, score, precision, recall)\n# test_ds = get_test_dataset(ordered=True) # since we are splitting the dataset and iterating separately on images and ids, order matters.\ntest_ds = get_test_dataset(ordered=True)\npredictions = predict(test_ds, model)\ntest_ids_ds = test_ds.map(lambda image, idnum: idnum).unbatch()\ntest_ids = next(iter(test_ids_ds.batch(NUM_TEST_IMAGES))).numpy().astype('U') # all in one batch\nsub_df = pd.DataFrame({'id': test_ids, 'label': predictions})\nsub_df.to_csv('submission.csv', index=False)","metadata":{"trusted":true},"execution_count":null,"outputs":[]}]}