{"cells":[{"metadata":{},"cell_type":"markdown","source":"# **TFRecord - Progressive Sprinkles**\n\n* In this Notebook I'll show how to use Progressive Sprinkles ([Read here](https://medium.com/@lessw/progressive-sprinkles-a-new-data-augmentation-for-cnns-and-helps-achieve-new-98-nih-malaria-6056965f671a)) with TFRecord.\n* This notebook forked from [AgentAuers](https://www.kaggle.com/agentauers) ([Link](https://www.kaggle.com/agentauers/incredible-tpus-finetune-effnetb0-b6-at-once)).\n* Progressive sprinkles implementation came from this [repository](https://github.com/Engineero/tf_sprinkles).\n* You can use a custom callback to change sprinkles configuration during training.\n* You can experiment with different sprinkles modes, normal, salt&pepper and gaussian:\n\n# Have fun!","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"DEVICE = \"TPU\"\n\nCFG = dict(\n    net_count         =   7,\n    batch_size        =  16,\n    \n    read_size         = 256, \n    crop_size         = 250, \n    net_size          = 224, \n    \n    LR_START          =   0.000005,\n    LR_MAX            =   0.000020,\n    LR_MIN            =   0.000001,\n    LR_RAMPUP_EPOCHS  =   5,\n    LR_SUSTAIN_EPOCHS =   0,\n    LR_EXP_DECAY      =   0.8,\n    epochs            =   5,\n    \n    rot               = 180.0,\n    shr               =   2.0,\n    hzoom             =   8.0,\n    wzoom             =   8.0,\n    hshift            =   8.0,\n    wshift            =   8.0,\n    \n    sprinkles_mode    = 'normal',\n    sprinkles_prob    =   1,\n    num_holes         =   40, \n    side_length       =   15,\n\n    optimizer         = 'adam',\n    label_smooth_fac  =   0.05,\n    \n    tta_steps         =  25    \n)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"!pip install -q efficientnet","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"import os, random, re, math, time\nrandom.seed(a=42)\n\nimport numpy as np\nimport pandas as pd\n\nimport tensorflow as tf\nimport tensorflow.keras.backend as K\nimport efficientnet.tfkeras as efn\n\nimport PIL\n\nfrom kaggle_datasets import KaggleDatasets\n\nfrom tqdm import tqdm","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"BASEPATH = \"../input/siim-isic-melanoma-classification\"\ndf_train = pd.read_csv(os.path.join(BASEPATH, 'train.csv'))\ndf_test  = pd.read_csv(os.path.join(BASEPATH, 'test.csv'))\ndf_sub   = pd.read_csv(os.path.join(BASEPATH, 'sample_submission.csv'))\n\nGCS_PATH    = KaggleDatasets().get_gcs_path('melanoma-256x256')\nfiles_train = np.sort(np.array(tf.io.gfile.glob(GCS_PATH + '/train*.tfrec')))\nfiles_test  = np.sort(np.array(tf.io.gfile.glob(GCS_PATH + '/test*.tfrec')))","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","trusted":true},"cell_type":"code","source":"if DEVICE == \"TPU\":\n    print(\"connecting to TPU...\")\n    try:\n        tpu = tf.distribute.cluster_resolver.TPUClusterResolver()\n        print('Running on TPU ', tpu.master())\n    except ValueError:\n        print(\"Could not connect to TPU\")\n        tpu = None\n\n    if tpu:\n        try:\n            print(\"initializing  TPU ...\")\n            tf.config.experimental_connect_to_cluster(tpu)\n            tf.tpu.experimental.initialize_tpu_system(tpu)\n            strategy = tf.distribute.experimental.TPUStrategy(tpu)\n            print(\"TPU initialized\")\n        except _:\n            print(\"failed to initialize TPU\")\n    else:\n        DEVICE = \"GPU\"\n\nif DEVICE != \"TPU\":\n    print(\"Using default strategy for CPU and single GPU\")\n    strategy = tf.distribute.get_strategy()\n\nif DEVICE == \"GPU\":\n    print(\"Num GPUs Available: \", len(tf.config.experimental.list_physical_devices('GPU')))\n    \n\nAUTO     = tf.data.experimental.AUTOTUNE\nREPLICAS = strategy.num_replicas_in_sync\nprint(f'REPLICAS: {REPLICAS}')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def get_mat(rotation, shear, height_zoom, width_zoom, height_shift, width_shift):\n    # returns 3x3 transformmatrix which transforms indicies\n        \n    # CONVERT DEGREES TO RADIANS\n    rotation = math.pi * rotation / 180.\n    shear    = math.pi * shear    / 180.\n\n    def get_3x3_mat(lst):\n        return tf.reshape(tf.concat([lst],axis=0), [3,3])\n    \n    # ROTATION MATRIX\n    c1   = tf.math.cos(rotation)\n    s1   = tf.math.sin(rotation)\n    one  = tf.constant([1],dtype='float32')\n    zero = tf.constant([0],dtype='float32')\n    \n    rotation_matrix = get_3x3_mat([c1,   s1,   zero, \n                                   -s1,  c1,   zero, \n                                   zero, zero, one])    \n    # SHEAR MATRIX\n    c2 = tf.math.cos(shear)\n    s2 = tf.math.sin(shear)    \n    \n    shear_matrix = get_3x3_mat([one,  s2,   zero, \n                                zero, c2,   zero, \n                                zero, zero, one])        \n    # ZOOM MATRIX\n    zoom_matrix = get_3x3_mat([one/height_zoom, zero,           zero, \n                               zero,            one/width_zoom, zero, \n                               zero,            zero,           one])    \n    # SHIFT MATRIX\n    shift_matrix = get_3x3_mat([one,  zero, height_shift, \n                                zero, one,  width_shift, \n                                zero, zero, one])\n    \n    return K.dot(K.dot(rotation_matrix, shear_matrix), \n                 K.dot(zoom_matrix,     shift_matrix))\n\n\ndef transform(image, cfg):    \n    # input image - is one image of size [dim,dim,3] not a batch of [b,dim,dim,3]\n    # output - image randomly rotated, sheared, zoomed, and shifted\n    DIM = cfg[\"read_size\"]\n    XDIM = DIM%2 #fix for size 331\n    \n    rot = cfg['rot'] * tf.random.normal([1], dtype='float32')\n    shr = cfg['shr'] * tf.random.normal([1], dtype='float32') \n    h_zoom = 1.0 + tf.random.normal([1], dtype='float32') / cfg['hzoom']\n    w_zoom = 1.0 + tf.random.normal([1], dtype='float32') / cfg['wzoom']\n    h_shift = cfg['hshift'] * tf.random.normal([1], dtype='float32') \n    w_shift = cfg['wshift'] * tf.random.normal([1], dtype='float32') \n\n    # GET TRANSFORMATION MATRIX\n    m = get_mat(rot,shr,h_zoom,w_zoom,h_shift,w_shift) \n\n    # LIST DESTINATION PIXEL INDICES\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    \n    # ROTATE DESTINATION PIXELS ONTO ORIGIN PIXELS\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    \n    # FIND ORIGIN PIXEL VALUES           \n    idx3 = tf.stack([DIM//2-idx2[0,], DIM//2-1+idx2[1,]])\n    d    = tf.gather_nd(image, tf.transpose(idx3))\n        \n    return tf.reshape(d,[DIM, DIM,3])","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Progressive Sprinkles Implementation","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"def make_mask(num_holes,side_length,rows, cols, num_channels):\n        \"\"\"Builds the mask for all sprinkles.\"\"\"\n        row_range = tf.tile(tf.range(rows)[..., tf.newaxis], [1, num_holes])\n        col_range = tf.tile(tf.range(cols)[..., tf.newaxis], [1, num_holes])\n        r_idx = tf.random.uniform([num_holes], minval=0, maxval=rows-1,\n                                  dtype=tf.int32)\n        c_idx = tf.random.uniform([num_holes], minval=0, maxval=cols-1,\n                                  dtype=tf.int32)\n        r1 = tf.clip_by_value(r_idx - side_length // 2, 0, rows)\n        r2 = tf.clip_by_value(r_idx + side_length // 2, 0, rows)\n        c1 = tf.clip_by_value(c_idx - side_length // 2, 0, cols)\n        c2 = tf.clip_by_value(c_idx + side_length // 2, 0, cols)\n        row_mask = (row_range > r1) & (row_range < r2)\n        col_mask = (col_range > c1) & (col_range < c2)\n\n        # Combine masks into one layer and duplicate over channels.\n        mask = row_mask[:, tf.newaxis] & col_mask\n        mask = tf.reduce_any(mask, axis=-1)\n        mask = mask[..., tf.newaxis]\n        mask = tf.tile(mask, [1, 1, num_channels])\n        return mask","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def sprinkles(image, cfg = CFG): \n    num_holes = cfg['num_holes']\n    side_length = cfg['side_length']\n    mode = cfg['sprinkles_mode']\n    PROBABILITY = cfg['sprinkles_prob']\n    \n    RandProb = tf.cast( tf.random.uniform([],0,1) < PROBABILITY, tf.int32)\n    if (RandProb == 0)|(num_holes == 0): return image\n    \n    img_shape = tf.shape(image)\n    if mode is 'normal':\n        rejected = tf.zeros_like(image)\n    elif mode is 'salt_pepper':\n        num_holes = num_holes // 2\n        rejected_high = tf.ones_like(image)\n        rejected_low = tf.zeros_like(image)\n    elif mode is 'gaussian':\n        rejected = tf.random.normal(img_shape, dtype=tf.float32)\n    else:\n        raise ValueError(f'Unknown mode \"{mode}\" given.')\n        \n    rows = img_shape[0]\n    cols = img_shape[1]\n    num_channels = img_shape[-1]\n    if mode is 'salt_pepper':\n        mask1 = make_mask(num_holes,side_length,rows, cols, num_channels)\n        mask2 = make_mask(num_holes,side_length,rows, cols, num_channels)\n        filtered_image = tf.where(mask1, rejected_high, image)\n        filtered_image = tf.where(mask2, rejected_low, filtered_image)\n    else:\n        mask = make_mask(num_holes,side_length,rows, cols, num_channels)\n        filtered_image = tf.where(mask, rejected, image)\n    return filtered_image","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def read_labeled_tfrecord(example):\n    tfrec_format = {\n        'image'                        : tf.io.FixedLenFeature([], tf.string),\n        'image_name'                   : tf.io.FixedLenFeature([], tf.string),\n        'patient_id'                   : tf.io.FixedLenFeature([], tf.int64),\n        'sex'                          : tf.io.FixedLenFeature([], tf.int64),\n        'age_approx'                   : tf.io.FixedLenFeature([], tf.int64),\n        'anatom_site_general_challenge': tf.io.FixedLenFeature([], tf.int64),\n        'diagnosis'                    : tf.io.FixedLenFeature([], tf.int64),\n        'target'                       : tf.io.FixedLenFeature([], tf.int64)\n    }           \n    example = tf.io.parse_single_example(example, tfrec_format)\n    return example['image'], example['target']\n\n\ndef read_unlabeled_tfrecord(example, return_image_name):\n    tfrec_format = {\n        'image'                        : tf.io.FixedLenFeature([], tf.string),\n        'image_name'                   : tf.io.FixedLenFeature([], tf.string),\n    }\n    example = tf.io.parse_single_example(example, tfrec_format)\n    return example['image'], example['image_name'] if return_image_name else 0\n\n \ndef prepare_image(img, cfg=None, augment=True):    \n    img = tf.image.decode_jpeg(img, channels=3)\n    img = tf.image.resize(img, [cfg['read_size'], cfg['read_size']])\n    img = tf.cast(img, tf.float32) / 255.0\n    \n    if augment:\n        img = transform(img, cfg)\n        img = sprinkles(img)\n        img = tf.image.random_crop(img, [cfg['crop_size'], cfg['crop_size'], 3])\n        img = tf.image.random_flip_left_right(img)\n       # img = tf.image.random_hue(img, 0.01)\n        img = tf.image.random_saturation(img, 0.7, 1.3)\n        img = tf.image.random_contrast(img, 0.8, 1.2)\n        img = tf.image.random_brightness(img, 0.1)\n\n    else:\n        img = tf.image.central_crop(img, cfg['crop_size'] / cfg['read_size'])\n                                   \n    img = tf.image.resize(img, [cfg['net_size'], cfg['net_size']])\n    img = tf.reshape(img, [cfg['net_size'], cfg['net_size'], 3])\n    return img\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)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def get_dataset(files, cfg, augment = True, shuffle = False, repeat = False, \n                labeled=True, return_image_names=True):\n    \n    ds = tf.data.TFRecordDataset(files, num_parallel_reads=AUTO)\n    ds = ds.cache()\n    \n    if repeat:\n        ds = ds.repeat()\n    \n    if shuffle: \n        ds = ds.shuffle(1024*8)\n        opt = tf.data.Options()\n        opt.experimental_deterministic = False\n        ds = ds.with_options(opt)\n        \n    if labeled: \n        ds = ds.map(read_labeled_tfrecord, num_parallel_calls=AUTO)\n    else:\n        ds = ds.map(lambda example: read_unlabeled_tfrecord(example, return_image_names), \n                    num_parallel_calls=AUTO)      \n    \n    ds = ds.map(lambda img, imgname_or_label: (prepare_image(img, augment=augment, cfg=cfg), \n                                               imgname_or_label), \n                num_parallel_calls=AUTO)\n    \n    ds = ds.batch(cfg['batch_size'] * REPLICAS)\n    ds = ds.prefetch(AUTO)\n    return ds\n\ndef show_dataset(thumb_size, cols, rows, ds):\n    mosaic = PIL.Image.new(mode='RGB', size=(thumb_size*cols + (cols-1), \n                                             thumb_size*rows + (rows-1)))\n   \n    for idx, data in enumerate(iter(ds)):\n        img, target_or_imgid = data\n        ix  = idx % cols\n        iy  = idx // cols\n        img = np.clip(img.numpy() * 255, 0, 255).astype(np.uint8)\n        img = PIL.Image.fromarray(img)\n        img = img.resize((thumb_size, thumb_size), resample=PIL.Image.BILINEAR)\n        mosaic.paste(img, (ix*thumb_size + ix, \n                           iy*thumb_size + iy))\n\n    display(mosaic)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"## Normal Progressive Sprinkles","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"ds = get_dataset(files_train, CFG).unbatch().take(6*5)   \nshow_dataset(128, 6, 5, ds)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"## Salt and Pepper Progressive Sprinkles","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"CFG['sprinkles_mode'] = 'salt_pepper'\n    \nds = get_dataset(files_train, CFG).unbatch().take(6*5)   \nshow_dataset(128, 6, 5, ds)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"## Gaussian Progressive Sprinkles","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"CFG['sprinkles_mode'] = 'gaussian'\n    \nds = get_dataset(files_train, CFG).unbatch().take(6*5)   \nshow_dataset(128, 6, 5, ds)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Custom Callback\n\n# Increase sprinkles number / size / probability during training","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"class IncreaseSprinklesHoles(tf.keras.callbacks.Callback):\n    def on_epoch_end(self, epoch, logs={}):\n        CFG['sprinkles_prob'] = epoch * 0.1\n        CFG['num_holes'] = epoch * 5\n        CFG['side_length'] = epoch * 3\n        \nsprinkles_cb = IncreaseSprinklesHoles()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"## Enjoy!","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"def get_lr_callback(cfg):\n    lr_start   = cfg['LR_START']\n    lr_max     = cfg['LR_MAX'] * strategy.num_replicas_in_sync\n    lr_min     = cfg['LR_MIN']\n    lr_ramp_ep = cfg['LR_RAMPUP_EPOCHS']\n    lr_sus_ep  = cfg['LR_SUSTAIN_EPOCHS']\n    lr_decay   = cfg['LR_EXP_DECAY']\n   \n    def lrfn(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    lr_callback = tf.keras.callbacks.LearningRateScheduler(lrfn, verbose=False)\n    return lr_callback","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def get_model(cfg):\n    model_input = tf.keras.Input(shape=(cfg['net_size'], cfg['net_size'], 3), name='imgIn')\n\n    dummy = tf.keras.layers.Lambda(lambda x:x)(model_input)\n    \n    outputs = []    \n    for i in range(cfg['net_count']):\n        constructor = getattr(efn, f'EfficientNetB{i}')\n        \n        x = constructor(include_top=False, weights='imagenet', \n                        input_shape=(cfg['net_size'], cfg['net_size'], 3), \n                        pooling='avg')(dummy)\n        \n        x = tf.keras.layers.Dense(1, activation='sigmoid')(x)\n        outputs.append(x)\n        \n    model = tf.keras.Model(model_input, outputs, name='aNetwork')\n    model.summary()\n    return model","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def compile_new_model(cfg):    \n    with strategy.scope():\n        model = get_model(cfg)\n     \n        losses = [tf.keras.losses.BinaryCrossentropy(label_smoothing = cfg['label_smooth_fac'])\n                  for i in range(cfg['net_count'])]\n        \n        model.compile(\n            optimizer = cfg['optimizer'],\n            loss      = losses,\n            metrics   = [tf.keras.metrics.AUC(name='auc')])\n        \n    return model","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"## Skip fit and predict","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"#ds_train     = get_dataset(files_train, CFG, augment=True, shuffle=True, repeat=True)\n#ds_train     = ds_train.map(lambda img, label: (img, tuple([label] * CFG['net_count'])))\n#\n#steps_train  = count_data_items(files_train) / (CFG['batch_size'] * REPLICAS)\n\n#model        = compile_new_model(CFG)\n#history      = model.fit(ds_train, \n#                         verbose          = 1,\n#                         steps_per_epoch  = steps_train, \n#                         epochs           = CFG['epochs'],\n#                         callbacks        = [get_lr_callback(CFG)])","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### predict the test set using augmented images","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"#CFG['batch_size'] = 256\n#\n#cnt_test   = count_data_items(files_test)\n#steps      = cnt_test / (CFG['batch_size'] * REPLICAS) * CFG['tta_steps']\n#ds_testAug = get_dataset(files_test, CFG, augment=True, repeat=True, \n#                         labeled=False, return_image_names=False)\n#\n#probs = model.predict(ds_testAug, verbose=1, steps=steps)\n#\n#probs = np.stack(probs)\n#probs = probs[:,:cnt_test * CFG['tta_steps']]\n#probs = np.stack(np.split(probs, CFG['tta_steps'], axis=1), axis=1)\n#probs = np.mean(probs, axis=1)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### sort predictions to have the same order as the submission\nThe submission ist sorted by image_name, but the dataset yielded a different order.\nTraverse the test dataset once again and capture the image_names. Then join this list of image_names with the predictions and sort by image_name.","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"#ds = get_dataset(files_test, CFG, augment=False, repeat=False, \n#                 labeled=False, return_image_names=True)\n#\n#image_names = np.array([img_name.numpy().decode(\"utf-8\") \n#                        for img, img_name in iter(ds.unbatch())])","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### write a submission file for each submodel","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"#for i in range(CFG[\"net_count\"]):\n#    submission = pd.DataFrame(dict(\n#        image_name = image_names,\n#        target     = probs[i,:,0]))\n#\n#    submission = submission.sort_values('image_name') \n#    submission.to_csv(f'submission_model_{i}.csv', index=False)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### write a submission file using the mean of all submodels","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"#submission = pd.DataFrame(dict(\n#    image_name = image_names,\n#    target     = np.mean(probs[:,:,0], axis=0)))\n#\n#submission = submission.sort_values('image_name') \n#submission.to_csv('submission_models_blended.csv', index=False)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#!ls -l .","execution_count":null,"outputs":[]}],"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":4,"nbformat_minor":4}