DeepCell_demo_train_pred_cyto.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Training with a Demo Cyto Dataset**<br>\n",
    "This notebook is based on Intro to DeepCell notebook. It works on tensorflow 2.7.1.<br>\n",
    "It trains a model with three heads: inner distance, outer distance, and fgbg.\n",
    "The dataset comprises images of 256x256."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.7.1\n"
     ]
    }
   ],
   "source": [
    "import syotil\n",
    "\n",
    "import numpy as np\n",
    "from skimage import io\n",
    "from matplotlib import pyplot as plt\n",
    "%matplotlib inline\n",
    "from timeit import default_timer\n",
    "\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.optimizers import SGD, Adam\n",
    "from tensorflow.keras.losses import MSE\n",
    "\n",
    "import deepcell\n",
    "from deepcell.utils.train_utils import rate_scheduler, get_callbacks, count_gpus\n",
    "from deepcell.losses import weighted_categorical_crossentropy\n",
    "from deepcell.model_zoo.panopticnet import PanopticNet\n",
    "from deepcell_toolbox.deep_watershed import deep_watershed\n",
    "from deepcell import image_generators\n",
    "\n",
    "print(tf.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "base_url = ('https://deepcell-data.s3-us-west-1.amazonaws.com/'\n",
    "            'demos/janelia/hela_s3_{}_256.npz')\n",
    "\n",
    "nuclear = deepcell.datasets.Dataset(\n",
    "    path='nuclear.npz',\n",
    "    url=base_url.format('nuclear'),\n",
    "     file_hash=None,\n",
    "    metadata=None\n",
    ")\n",
    "\n",
    "fluo_cyto = deepcell.datasets.Dataset(\n",
    "    path='fluo_cyto.npz',\n",
    "    url=base_url.format('fluo_cyto'),\n",
    "     file_hash=None,\n",
    "    metadata=None\n",
    ")\n",
    "\n",
    "phase = deepcell.datasets.Dataset(\n",
    "    path='phase.npz',\n",
    "    url=base_url.format('phase'),\n",
    "     file_hash=None,\n",
    "    metadata=None\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train.shape: (1872, 256, 256, 1)\n",
      "X_test.shape: (468, 256, 256, 1)\n"
     ]
    }
   ],
   "source": [
    "dataset = fluo_cyto  # nuclear, phase, fluo_cyto\n",
    "model_name = 'cyto_deep_watershed'\n",
    "model_path = '{}.h5'.format(model_name)\n",
    "\n",
    "min_objects = 2\n",
    "\n",
    "test_size = 0.2 # fraction of data saved as test\n",
    "seed = 0 # seed for random train-test split\n",
    "\n",
    "(X_train, y_train), (X_test, y_test) = dataset.load_data(\n",
    "    test_size=test_size, seed=seed)\n",
    "\n",
    "print('X_train.shape: {}\\nX_test.shape: {}'.format(\n",
    "    X_train.shape, X_test.shape))\n",
    "transforms = ['inner-distance', 'outer-distance', 'fgbg']\n",
    "transforms_kwargs = {'outer-distance': {'erosion_width': 0}}\n",
    "\n",
    "\n",
    "# use augmentation for training but not validation\n",
    "datagen = image_generators.SemanticDataGenerator(\n",
    "    rotation_range=180,\n",
    "    fill_mode='reflect',\n",
    "    zoom_range=(0.75, 1.25),\n",
    "    horizontal_flip=True,\n",
    "    vertical_flip=True)\n",
    "\n",
    "datagen_val = image_generators.SemanticDataGenerator()\n",
    "\n",
    "batch_size = 8\n",
    "\n",
    "train_data = datagen.flow(\n",
    "    {'X': X_train, 'y': y_train},\n",
    "    seed=seed,\n",
    "    transforms=transforms,\n",
    "    transforms_kwargs=transforms_kwargs,\n",
    "    min_objects=min_objects,\n",
    "    batch_size=batch_size)\n",
    "\n",
    "val_data = datagen_val.flow(\n",
    "    {'X': X_test, 'y': y_test},\n",
    "    seed=seed,\n",
    "    transforms=transforms,\n",
    "    transforms_kwargs=transforms_kwargs,\n",
    "    min_objects=min_objects,\n",
    "    batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1872, 256, 256, 1)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/app/software/scikit-image/0.19.1-foss-2021b/lib/python3.9/site-packages/skimage/io/_plugins/matplotlib_plugin.py:150: UserWarning: Low image data range; displaying image with stretched contrast.\n",
      "  lo, hi, cmap = _get_display_range(image)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(y_train.shape)\n",
    "\n",
    "i=5\n",
    "\n",
    "io.imshow(X_train[i,:,:,0])\n",
    "plt.show()\n",
    "#print(X_train[i,:,:,0])\n",
    "\n",
    "tmp = syotil.masks_to_outlines(y_train[i,:,:,0])\n",
    "io.imshow(tmp)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-10-31 10:24:34.928414: W tensorflow/python/util/util.cc:368] Sets are not currently considered sequences, but this may change in the future, so consider avoiding using them.\n",
      "2022-10-31 10:24:37.408082: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1525] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 10429 MB memory:  -> device: 0, name: NVIDIA GeForce GTX 1080 Ti, pci bus id: 0000:5e:00.0, compute capability: 6.1\n",
      "/app/software/TensorFlow/2.7.1-foss-2021b-CUDA-11.4.1/lib/python3.9/site-packages/keras/optimizer_v2/adam.py:105: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
      "  super(Adam, self).__init__(name, **kwargs)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[('semantic_0', (None, 256, 256, 1)),\n",
       " ('semantic_1', (None, 256, 256, 1)),\n",
       " ('semantic_2', (None, 256, 256, 2))]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "semantic_classes = [1, 1, 2] # inner distance, outer distance, fgbg\n",
    "\n",
    "model = PanopticNet(\n",
    "    backbone='resnet50',\n",
    "    input_shape=X_train.shape[1:],\n",
    "    norm_method='whole_image',\n",
    "    num_semantic_classes=semantic_classes)\n",
    "\n",
    "lr = 1e-4\n",
    "optimizer = Adam(lr=lr, clipnorm=0.001)\n",
    "lr_sched = rate_scheduler(lr=lr, decay=0.99)\n",
    "\n",
    "# Create a dictionary of losses for each semantic head\n",
    "\n",
    "def semantic_loss(n_classes):\n",
    "    def _semantic_loss(y_pred, y_true):\n",
    "        if n_classes > 1:\n",
    "            return 0.01 * weighted_categorical_crossentropy(\n",
    "                y_pred, y_true, n_classes=n_classes)\n",
    "        return MSE(y_pred, y_true)\n",
    "    return _semantic_loss\n",
    "\n",
    "loss = {}\n",
    "\n",
    "# Give losses for all of the semantic heads\n",
    "for layer in model.layers:\n",
    "    if layer.name.startswith('semantic_'):\n",
    "        n_classes = layer.output_shape[-1]\n",
    "        loss[layer.name] = semantic_loss(n_classes)\n",
    "        \n",
    "model.compile(loss=loss, optimizer=optimizer)\n",
    "\n",
    "[(layer.name, layer.output_shape) for layer in filter(lambda x: x.name.startswith('semantic_'), model.layers)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-10-31 10:24:42.191405: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1525] Created device /device:GPU:0 with 10429 MB memory:  -> device: 0, name: NVIDIA GeForce GTX 1080 Ti, pci bus id: 0000:5e:00.0, compute capability: 6.1\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training on 1 GPUs.\n",
      "Epoch 1/50\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-10-31 10:25:00.761305: I tensorflow/stream_executor/cuda/cuda_dnn.cc:366] Loaded cuDNN version 8202\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "234/234 [==============================] - ETA: 0s - loss: 0.0494 - semantic_0_loss: 0.0209 - semantic_1_loss: 0.0268 - semantic_2_loss: 0.0017\n",
      "Epoch 00001: val_loss improved from inf to 0.07980, saving model to cyto_deep_watershed.h5\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/app/software/TensorFlow/2.7.1-foss-2021b-CUDA-11.4.1/lib/python3.9/site-packages/keras/engine/functional.py:1410: CustomMaskWarning: Custom mask layers require a config and must override get_config. When loading, the custom mask layer must be passed to the custom_objects argument.\n",
      "  layer_config = serialize_layer_fn(layer)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "234/234 [==============================] - 100s 287ms/step - loss: 0.0494 - semantic_0_loss: 0.0209 - semantic_1_loss: 0.0268 - semantic_2_loss: 0.0017 - val_loss: 0.0798 - val_semantic_0_loss: 0.0045 - val_semantic_1_loss: 0.0566 - val_semantic_2_loss: 0.0187 - lr: 1.0000e-04\n",
      "Epoch 2/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0091 - semantic_0_loss: 0.0022 - semantic_1_loss: 0.0062 - semantic_2_loss: 8.0713e-04\n",
      "Epoch 00002: val_loss improved from 0.07980 to 0.01308, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 64s 271ms/step - loss: 0.0091 - semantic_0_loss: 0.0022 - semantic_1_loss: 0.0062 - semantic_2_loss: 8.0713e-04 - val_loss: 0.0131 - val_semantic_0_loss: 0.0027 - val_semantic_1_loss: 0.0090 - val_semantic_2_loss: 0.0014 - lr: 9.9000e-05\n",
      "Epoch 3/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0077 - semantic_0_loss: 0.0019 - semantic_1_loss: 0.0051 - semantic_2_loss: 6.8774e-04\n",
      "Epoch 00003: val_loss improved from 0.01308 to 0.00898, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 64s 273ms/step - loss: 0.0077 - semantic_0_loss: 0.0019 - semantic_1_loss: 0.0051 - semantic_2_loss: 6.8774e-04 - val_loss: 0.0090 - val_semantic_0_loss: 0.0022 - val_semantic_1_loss: 0.0059 - val_semantic_2_loss: 8.1039e-04 - lr: 9.8010e-05\n",
      "Epoch 4/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0071 - semantic_0_loss: 0.0018 - semantic_1_loss: 0.0046 - semantic_2_loss: 6.5158e-04\n",
      "Epoch 00004: val_loss improved from 0.00898 to 0.00784, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 64s 272ms/step - loss: 0.0071 - semantic_0_loss: 0.0018 - semantic_1_loss: 0.0046 - semantic_2_loss: 6.5158e-04 - val_loss: 0.0078 - val_semantic_0_loss: 0.0026 - val_semantic_1_loss: 0.0046 - val_semantic_2_loss: 6.2407e-04 - lr: 9.7030e-05\n",
      "Epoch 5/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0064 - semantic_0_loss: 0.0017 - semantic_1_loss: 0.0041 - semantic_2_loss: 6.2583e-04\n",
      "Epoch 00005: val_loss improved from 0.00784 to 0.00646, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 63s 270ms/step - loss: 0.0064 - semantic_0_loss: 0.0017 - semantic_1_loss: 0.0041 - semantic_2_loss: 6.2583e-04 - val_loss: 0.0065 - val_semantic_0_loss: 0.0019 - val_semantic_1_loss: 0.0039 - val_semantic_2_loss: 6.5052e-04 - lr: 9.6060e-05\n",
      "Epoch 6/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0059 - semantic_0_loss: 0.0016 - semantic_1_loss: 0.0037 - semantic_2_loss: 5.9897e-04\n",
      "Epoch 00006: val_loss improved from 0.00646 to 0.00645, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 64s 272ms/step - loss: 0.0059 - semantic_0_loss: 0.0016 - semantic_1_loss: 0.0037 - semantic_2_loss: 5.9897e-04 - val_loss: 0.0064 - val_semantic_0_loss: 0.0019 - val_semantic_1_loss: 0.0039 - val_semantic_2_loss: 5.9654e-04 - lr: 9.5099e-05\n",
      "Epoch 7/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0056 - semantic_0_loss: 0.0015 - semantic_1_loss: 0.0034 - semantic_2_loss: 5.8409e-04\n",
      "Epoch 00007: val_loss improved from 0.00645 to 0.00605, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 63s 270ms/step - loss: 0.0056 - semantic_0_loss: 0.0015 - semantic_1_loss: 0.0034 - semantic_2_loss: 5.8409e-04 - val_loss: 0.0061 - val_semantic_0_loss: 0.0018 - val_semantic_1_loss: 0.0036 - val_semantic_2_loss: 5.8553e-04 - lr: 9.4148e-05\n",
      "Epoch 8/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0054 - semantic_0_loss: 0.0015 - semantic_1_loss: 0.0033 - semantic_2_loss: 5.7193e-04\n",
      "Epoch 00008: val_loss improved from 0.00605 to 0.00558, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 64s 272ms/step - loss: 0.0054 - semantic_0_loss: 0.0015 - semantic_1_loss: 0.0033 - semantic_2_loss: 5.7193e-04 - val_loss: 0.0056 - val_semantic_0_loss: 0.0018 - val_semantic_1_loss: 0.0032 - val_semantic_2_loss: 5.7371e-04 - lr: 9.3207e-05\n",
      "Epoch 9/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0052 - semantic_0_loss: 0.0015 - semantic_1_loss: 0.0032 - semantic_2_loss: 5.6675e-04\n",
      "Epoch 00009: val_loss did not improve from 0.00558\n",
      "234/234 [==============================] - 63s 267ms/step - loss: 0.0052 - semantic_0_loss: 0.0015 - semantic_1_loss: 0.0032 - semantic_2_loss: 5.6675e-04 - val_loss: 0.0058 - val_semantic_0_loss: 0.0019 - val_semantic_1_loss: 0.0034 - val_semantic_2_loss: 5.6850e-04 - lr: 9.2274e-05\n",
      "Epoch 10/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0051 - semantic_0_loss: 0.0014 - semantic_1_loss: 0.0031 - semantic_2_loss: 5.6045e-04\n",
      "Epoch 00010: val_loss did not improve from 0.00558\n",
      "234/234 [==============================] - 62s 264ms/step - loss: 0.0051 - semantic_0_loss: 0.0014 - semantic_1_loss: 0.0031 - semantic_2_loss: 5.6045e-04 - val_loss: 0.0056 - val_semantic_0_loss: 0.0018 - val_semantic_1_loss: 0.0033 - val_semantic_2_loss: 5.4568e-04 - lr: 9.1352e-05\n",
      "Epoch 11/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0049 - semantic_0_loss: 0.0014 - semantic_1_loss: 0.0030 - semantic_2_loss: 5.4816e-04\n",
      "Epoch 00011: val_loss improved from 0.00558 to 0.00540, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 64s 274ms/step - loss: 0.0049 - semantic_0_loss: 0.0014 - semantic_1_loss: 0.0030 - semantic_2_loss: 5.4816e-04 - val_loss: 0.0054 - val_semantic_0_loss: 0.0018 - val_semantic_1_loss: 0.0031 - val_semantic_2_loss: 5.2746e-04 - lr: 9.0438e-05\n",
      "Epoch 12/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0048 - semantic_0_loss: 0.0014 - semantic_1_loss: 0.0029 - semantic_2_loss: 5.4710e-04\n",
      "Epoch 00012: val_loss did not improve from 0.00540\n",
      "234/234 [==============================] - 62s 265ms/step - loss: 0.0048 - semantic_0_loss: 0.0014 - semantic_1_loss: 0.0029 - semantic_2_loss: 5.4710e-04 - val_loss: 0.0055 - val_semantic_0_loss: 0.0018 - val_semantic_1_loss: 0.0031 - val_semantic_2_loss: 5.6900e-04 - lr: 8.9534e-05\n",
      "Epoch 13/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0047 - semantic_0_loss: 0.0014 - semantic_1_loss: 0.0028 - semantic_2_loss: 5.4221e-04\n",
      "Epoch 00013: val_loss did not improve from 0.00540\n",
      "234/234 [==============================] - 61s 258ms/step - loss: 0.0047 - semantic_0_loss: 0.0014 - semantic_1_loss: 0.0028 - semantic_2_loss: 5.4221e-04 - val_loss: 0.0054 - val_semantic_0_loss: 0.0018 - val_semantic_1_loss: 0.0030 - val_semantic_2_loss: 6.0794e-04 - lr: 8.8638e-05\n",
      "Epoch 14/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0046 - semantic_0_loss: 0.0013 - semantic_1_loss: 0.0028 - semantic_2_loss: 5.3317e-04\n",
      "Epoch 00014: val_loss improved from 0.00540 to 0.00511, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 64s 272ms/step - loss: 0.0046 - semantic_0_loss: 0.0013 - semantic_1_loss: 0.0028 - semantic_2_loss: 5.3317e-04 - val_loss: 0.0051 - val_semantic_0_loss: 0.0017 - val_semantic_1_loss: 0.0029 - val_semantic_2_loss: 5.1769e-04 - lr: 8.7752e-05\n",
      "Epoch 15/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0045 - semantic_0_loss: 0.0013 - semantic_1_loss: 0.0027 - semantic_2_loss: 5.3169e-04\n",
      "Epoch 00015: val_loss improved from 0.00511 to 0.00499, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 63s 270ms/step - loss: 0.0045 - semantic_0_loss: 0.0013 - semantic_1_loss: 0.0027 - semantic_2_loss: 5.3169e-04 - val_loss: 0.0050 - val_semantic_0_loss: 0.0017 - val_semantic_1_loss: 0.0028 - val_semantic_2_loss: 5.1060e-04 - lr: 8.6875e-05\n",
      "Epoch 16/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0044 - semantic_0_loss: 0.0013 - semantic_1_loss: 0.0026 - semantic_2_loss: 5.2811e-04\n",
      "Epoch 00016: val_loss did not improve from 0.00499\n",
      "234/234 [==============================] - 63s 266ms/step - loss: 0.0044 - semantic_0_loss: 0.0013 - semantic_1_loss: 0.0026 - semantic_2_loss: 5.2811e-04 - val_loss: 0.0052 - val_semantic_0_loss: 0.0018 - val_semantic_1_loss: 0.0029 - val_semantic_2_loss: 5.2595e-04 - lr: 8.6006e-05\n",
      "Epoch 17/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0043 - semantic_0_loss: 0.0013 - semantic_1_loss: 0.0026 - semantic_2_loss: 5.2003e-04\n",
      "Epoch 00017: val_loss did not improve from 0.00499\n",
      "234/234 [==============================] - 61s 260ms/step - loss: 0.0043 - semantic_0_loss: 0.0013 - semantic_1_loss: 0.0026 - semantic_2_loss: 5.2003e-04 - val_loss: 0.0050 - val_semantic_0_loss: 0.0017 - val_semantic_1_loss: 0.0028 - val_semantic_2_loss: 5.1590e-04 - lr: 8.5146e-05\n",
      "Epoch 18/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0042 - semantic_0_loss: 0.0012 - semantic_1_loss: 0.0025 - semantic_2_loss: 5.1774e-04\n",
      "Epoch 00018: val_loss did not improve from 0.00499\n",
      "234/234 [==============================] - 63s 266ms/step - loss: 0.0042 - semantic_0_loss: 0.0012 - semantic_1_loss: 0.0025 - semantic_2_loss: 5.1774e-04 - val_loss: 0.0050 - val_semantic_0_loss: 0.0017 - val_semantic_1_loss: 0.0028 - val_semantic_2_loss: 5.0587e-04 - lr: 8.4294e-05\n",
      "Epoch 19/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0042 - semantic_0_loss: 0.0012 - semantic_1_loss: 0.0025 - semantic_2_loss: 5.1285e-04\n",
      "Epoch 00019: val_loss did not improve from 0.00499\n",
      "234/234 [==============================] - 63s 268ms/step - loss: 0.0042 - semantic_0_loss: 0.0012 - semantic_1_loss: 0.0025 - semantic_2_loss: 5.1285e-04 - val_loss: 0.0050 - val_semantic_0_loss: 0.0017 - val_semantic_1_loss: 0.0028 - val_semantic_2_loss: 5.5047e-04 - lr: 8.3451e-05\n",
      "Epoch 20/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0041 - semantic_0_loss: 0.0012 - semantic_1_loss: 0.0024 - semantic_2_loss: 5.0407e-04\n",
      "Epoch 00020: val_loss improved from 0.00499 to 0.00494, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 66s 282ms/step - loss: 0.0041 - semantic_0_loss: 0.0012 - semantic_1_loss: 0.0024 - semantic_2_loss: 5.0407e-04 - val_loss: 0.0049 - val_semantic_0_loss: 0.0017 - val_semantic_1_loss: 0.0027 - val_semantic_2_loss: 5.1310e-04 - lr: 8.2617e-05\n",
      "Epoch 21/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0041 - semantic_0_loss: 0.0012 - semantic_1_loss: 0.0024 - semantic_2_loss: 5.0141e-04\n",
      "Epoch 00021: val_loss improved from 0.00494 to 0.00482, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 65s 279ms/step - loss: 0.0041 - semantic_0_loss: 0.0012 - semantic_1_loss: 0.0024 - semantic_2_loss: 5.0141e-04 - val_loss: 0.0048 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0027 - val_semantic_2_loss: 5.0410e-04 - lr: 8.1791e-05\n",
      "Epoch 22/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0040 - semantic_0_loss: 0.0011 - semantic_1_loss: 0.0024 - semantic_2_loss: 5.0072e-04\n",
      "Epoch 00022: val_loss improved from 0.00482 to 0.00468, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 65s 279ms/step - loss: 0.0040 - semantic_0_loss: 0.0011 - semantic_1_loss: 0.0024 - semantic_2_loss: 5.0072e-04 - val_loss: 0.0047 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0026 - val_semantic_2_loss: 4.9225e-04 - lr: 8.0973e-05\n",
      "Epoch 23/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0039 - semantic_0_loss: 0.0011 - semantic_1_loss: 0.0023 - semantic_2_loss: 4.9752e-04\n",
      "Epoch 00023: val_loss did not improve from 0.00468\n",
      "234/234 [==============================] - 61s 261ms/step - loss: 0.0039 - semantic_0_loss: 0.0011 - semantic_1_loss: 0.0023 - semantic_2_loss: 4.9752e-04 - val_loss: 0.0047 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0026 - val_semantic_2_loss: 4.8983e-04 - lr: 8.0163e-05\n",
      "Epoch 24/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0039 - semantic_0_loss: 0.0011 - semantic_1_loss: 0.0023 - semantic_2_loss: 4.9704e-04\n",
      "Epoch 00024: val_loss did not improve from 0.00468\n",
      "234/234 [==============================] - 63s 267ms/step - loss: 0.0039 - semantic_0_loss: 0.0011 - semantic_1_loss: 0.0023 - semantic_2_loss: 4.9704e-04 - val_loss: 0.0049 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0028 - val_semantic_2_loss: 5.0859e-04 - lr: 7.9361e-05\n",
      "Epoch 25/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0038 - semantic_0_loss: 0.0011 - semantic_1_loss: 0.0022 - semantic_2_loss: 4.9552e-04\n",
      "Epoch 00025: val_loss did not improve from 0.00468\n",
      "234/234 [==============================] - 62s 262ms/step - loss: 0.0038 - semantic_0_loss: 0.0011 - semantic_1_loss: 0.0022 - semantic_2_loss: 4.9552e-04 - val_loss: 0.0047 - val_semantic_0_loss: 0.0017 - val_semantic_1_loss: 0.0026 - val_semantic_2_loss: 4.8142e-04 - lr: 7.8568e-05\n",
      "Epoch 26/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0038 - semantic_0_loss: 0.0011 - semantic_1_loss: 0.0022 - semantic_2_loss: 4.9275e-04\n",
      "Epoch 00026: val_loss improved from 0.00468 to 0.00467, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 64s 272ms/step - loss: 0.0038 - semantic_0_loss: 0.0011 - semantic_1_loss: 0.0022 - semantic_2_loss: 4.9275e-04 - val_loss: 0.0047 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0026 - val_semantic_2_loss: 4.7575e-04 - lr: 7.7782e-05\n",
      "Epoch 27/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0037 - semantic_0_loss: 0.0011 - semantic_1_loss: 0.0022 - semantic_2_loss: 4.8333e-04\n",
      "Epoch 00027: val_loss did not improve from 0.00467\n",
      "234/234 [==============================] - 64s 272ms/step - loss: 0.0037 - semantic_0_loss: 0.0011 - semantic_1_loss: 0.0022 - semantic_2_loss: 4.8333e-04 - val_loss: 0.0047 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0026 - val_semantic_2_loss: 4.7200e-04 - lr: 7.7004e-05\n",
      "Epoch 28/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0037 - semantic_0_loss: 0.0010 - semantic_1_loss: 0.0022 - semantic_2_loss: 4.8835e-04\n",
      "Epoch 00028: val_loss did not improve from 0.00467\n",
      "234/234 [==============================] - 62s 263ms/step - loss: 0.0037 - semantic_0_loss: 0.0010 - semantic_1_loss: 0.0022 - semantic_2_loss: 4.8835e-04 - val_loss: 0.0047 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0026 - val_semantic_2_loss: 4.8610e-04 - lr: 7.6234e-05\n",
      "Epoch 29/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0036 - semantic_0_loss: 0.0010 - semantic_1_loss: 0.0021 - semantic_2_loss: 4.8431e-04\n",
      "Epoch 00029: val_loss did not improve from 0.00467\n",
      "234/234 [==============================] - 62s 262ms/step - loss: 0.0036 - semantic_0_loss: 0.0010 - semantic_1_loss: 0.0021 - semantic_2_loss: 4.8431e-04 - val_loss: 0.0048 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0026 - val_semantic_2_loss: 4.8837e-04 - lr: 7.5472e-05\n",
      "Epoch 30/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0036 - semantic_0_loss: 0.0010 - semantic_1_loss: 0.0021 - semantic_2_loss: 4.7830e-04\n",
      "Epoch 00030: val_loss did not improve from 0.00467\n",
      "234/234 [==============================] - 63s 267ms/step - loss: 0.0036 - semantic_0_loss: 0.0010 - semantic_1_loss: 0.0021 - semantic_2_loss: 4.7830e-04 - val_loss: 0.0047 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0026 - val_semantic_2_loss: 4.7287e-04 - lr: 7.4717e-05\n",
      "Epoch 31/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0035 - semantic_0_loss: 9.9665e-04 - semantic_1_loss: 0.0021 - semantic_2_loss: 4.7672e-04\n",
      "Epoch 00031: val_loss did not improve from 0.00467\n",
      "234/234 [==============================] - 62s 265ms/step - loss: 0.0035 - semantic_0_loss: 9.9665e-04 - semantic_1_loss: 0.0021 - semantic_2_loss: 4.7672e-04 - val_loss: 0.0047 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0026 - val_semantic_2_loss: 4.8504e-04 - lr: 7.3970e-05\n",
      "Epoch 32/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0035 - semantic_0_loss: 9.8696e-04 - semantic_1_loss: 0.0020 - semantic_2_loss: 4.7829e-04\n",
      "Epoch 00032: val_loss improved from 0.00467 to 0.00450, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 64s 274ms/step - loss: 0.0035 - semantic_0_loss: 9.8696e-04 - semantic_1_loss: 0.0020 - semantic_2_loss: 4.7829e-04 - val_loss: 0.0045 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0024 - val_semantic_2_loss: 4.7489e-04 - lr: 7.3230e-05\n",
      "Epoch 33/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "234/234 [==============================] - ETA: 0s - loss: 0.0035 - semantic_0_loss: 9.7155e-04 - semantic_1_loss: 0.0020 - semantic_2_loss: 4.7531e-04\n",
      "Epoch 00033: val_loss did not improve from 0.00450\n",
      "234/234 [==============================] - 62s 262ms/step - loss: 0.0035 - semantic_0_loss: 9.7155e-04 - semantic_1_loss: 0.0020 - semantic_2_loss: 4.7531e-04 - val_loss: 0.0047 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0027 - val_semantic_2_loss: 4.8741e-04 - lr: 7.2498e-05\n",
      "Epoch 34/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0034 - semantic_0_loss: 9.6043e-04 - semantic_1_loss: 0.0020 - semantic_2_loss: 4.7057e-04\n",
      "Epoch 00034: val_loss improved from 0.00450 to 0.00447, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 66s 281ms/step - loss: 0.0034 - semantic_0_loss: 9.6043e-04 - semantic_1_loss: 0.0020 - semantic_2_loss: 4.7057e-04 - val_loss: 0.0045 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0024 - val_semantic_2_loss: 4.6237e-04 - lr: 7.1773e-05\n",
      "Epoch 35/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0034 - semantic_0_loss: 9.4520e-04 - semantic_1_loss: 0.0020 - semantic_2_loss: 4.7275e-04\n",
      "Epoch 00035: val_loss did not improve from 0.00447\n",
      "234/234 [==============================] - 64s 273ms/step - loss: 0.0034 - semantic_0_loss: 9.4520e-04 - semantic_1_loss: 0.0020 - semantic_2_loss: 4.7275e-04 - val_loss: 0.0045 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0025 - val_semantic_2_loss: 4.6632e-04 - lr: 7.1055e-05\n",
      "Epoch 36/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0034 - semantic_0_loss: 9.5044e-04 - semantic_1_loss: 0.0020 - semantic_2_loss: 4.7163e-04\n",
      "Epoch 00036: val_loss improved from 0.00447 to 0.00445, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 66s 280ms/step - loss: 0.0034 - semantic_0_loss: 9.5044e-04 - semantic_1_loss: 0.0020 - semantic_2_loss: 4.7163e-04 - val_loss: 0.0044 - val_semantic_0_loss: 0.0015 - val_semantic_1_loss: 0.0025 - val_semantic_2_loss: 4.6822e-04 - lr: 7.0345e-05\n",
      "Epoch 37/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0033 - semantic_0_loss: 9.2160e-04 - semantic_1_loss: 0.0020 - semantic_2_loss: 4.6573e-04\n",
      "Epoch 00037: val_loss did not improve from 0.00445\n",
      "234/234 [==============================] - 62s 264ms/step - loss: 0.0033 - semantic_0_loss: 9.2160e-04 - semantic_1_loss: 0.0020 - semantic_2_loss: 4.6573e-04 - val_loss: 0.0045 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0025 - val_semantic_2_loss: 4.7379e-04 - lr: 6.9641e-05\n",
      "Epoch 38/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0033 - semantic_0_loss: 9.1330e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6514e-04\n",
      "Epoch 00038: val_loss improved from 0.00445 to 0.00439, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 63s 268ms/step - loss: 0.0033 - semantic_0_loss: 9.1330e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6514e-04 - val_loss: 0.0044 - val_semantic_0_loss: 0.0015 - val_semantic_1_loss: 0.0024 - val_semantic_2_loss: 4.6311e-04 - lr: 6.8945e-05\n",
      "Epoch 39/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0033 - semantic_0_loss: 9.3037e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6504e-04\n",
      "Epoch 00039: val_loss improved from 0.00439 to 0.00431, saving model to cyto_deep_watershed.h5\n",
      "234/234 [==============================] - 65s 275ms/step - loss: 0.0033 - semantic_0_loss: 9.3037e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6504e-04 - val_loss: 0.0043 - val_semantic_0_loss: 0.0015 - val_semantic_1_loss: 0.0024 - val_semantic_2_loss: 4.6283e-04 - lr: 6.8255e-05\n",
      "Epoch 40/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0033 - semantic_0_loss: 8.9607e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6104e-04\n",
      "Epoch 00040: val_loss did not improve from 0.00431\n",
      "234/234 [==============================] - 63s 268ms/step - loss: 0.0033 - semantic_0_loss: 8.9607e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6104e-04 - val_loss: 0.0044 - val_semantic_0_loss: 0.0015 - val_semantic_1_loss: 0.0024 - val_semantic_2_loss: 4.5483e-04 - lr: 6.7573e-05\n",
      "Epoch 41/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0032 - semantic_0_loss: 8.8459e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6045e-04\n",
      "Epoch 00041: val_loss did not improve from 0.00431\n",
      "234/234 [==============================] - 62s 266ms/step - loss: 0.0032 - semantic_0_loss: 8.8459e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6045e-04 - val_loss: 0.0044 - val_semantic_0_loss: 0.0015 - val_semantic_1_loss: 0.0024 - val_semantic_2_loss: 4.7365e-04 - lr: 6.6897e-05\n",
      "Epoch 42/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0032 - semantic_0_loss: 8.7253e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6160e-04\n",
      "Epoch 00042: val_loss did not improve from 0.00431\n",
      "234/234 [==============================] - 61s 258ms/step - loss: 0.0032 - semantic_0_loss: 8.7253e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6160e-04 - val_loss: 0.0045 - val_semantic_0_loss: 0.0016 - val_semantic_1_loss: 0.0025 - val_semantic_2_loss: 4.7877e-04 - lr: 6.6228e-05\n",
      "Epoch 43/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0032 - semantic_0_loss: 8.8229e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6185e-04\n",
      "Epoch 00043: val_loss did not improve from 0.00431\n",
      "234/234 [==============================] - 62s 266ms/step - loss: 0.0032 - semantic_0_loss: 8.8229e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6185e-04 - val_loss: 0.0043 - val_semantic_0_loss: 0.0015 - val_semantic_1_loss: 0.0024 - val_semantic_2_loss: 4.5971e-04 - lr: 6.5566e-05\n",
      "Epoch 44/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0032 - semantic_0_loss: 8.7709e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6255e-04\n",
      "Epoch 00044: val_loss did not improve from 0.00431\n",
      "234/234 [==============================] - 63s 269ms/step - loss: 0.0032 - semantic_0_loss: 8.7709e-04 - semantic_1_loss: 0.0019 - semantic_2_loss: 4.6255e-04 - val_loss: 0.0044 - val_semantic_0_loss: 0.0015 - val_semantic_1_loss: 0.0024 - val_semantic_2_loss: 4.6427e-04 - lr: 6.4910e-05\n",
      "Epoch 45/50\n",
      "234/234 [==============================] - ETA: 0s - loss: 0.0031 - semantic_0_loss: 8.5323e-04 - semantic_1_loss: 0.0018 - semantic_2_loss: 4.5826e-04\n",
      "Epoch 00045: val_loss did not improve from 0.00431\n",
      "234/234 [==============================] - 62s 263ms/step - loss: 0.0031 - semantic_0_loss: 8.5323e-04 - semantic_1_loss: 0.0018 - semantic_2_loss: 4.5826e-04 - val_loss: 0.0043 - val_semantic_0_loss: 0.0015 - val_semantic_1_loss: 0.0023 - val_semantic_2_loss: 4.5241e-04 - lr: 6.4261e-05\n",
      "Epoch 46/50\n",
      " 40/234 [====>.........................] - ETA: 46s - loss: 0.0032 - semantic_0_loss: 8.6657e-04 - semantic_1_loss: 0.0018 - semantic_2_loss: 4.6961e-04"
     ]
    }
   ],
   "source": [
    "# fit the model\n",
    "\n",
    "print('Training on', count_gpus(), 'GPUs.')\n",
    "\n",
    "train_callbacks = get_callbacks(\n",
    "    model_path,\n",
    "    lr_sched=lr_sched,\n",
    "    monitor='val_loss',\n",
    "    verbose=1)\n",
    "\n",
    "loss_history = model.fit(\n",
    "    train_data,\n",
    "    steps_per_epoch=train_data.y.shape[0] // batch_size,\n",
    "    epochs=50,\n",
    "    validation_data=val_data,\n",
    "    validation_steps=val_data.y.shape[0] // batch_size,\n",
    "    callbacks=train_callbacks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create prediction model\n",
    "prediction_model = PanopticNet(\n",
    "    backbone='resnet50',\n",
    "    norm_method='whole_image',\n",
    "    num_semantic_classes=[1, 1], # inner distance, outer distance\n",
    "    input_shape=X_test.shape[1:]\n",
    ")\n",
    "\n",
    "prediction_model.load_weights(model_path, by_name=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-10-31 11:22:39.374732: I tensorflow/stream_executor/cuda/cuda_dnn.cc:366] Loaded cuDNN version 8202\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Watershed segmentation of shape (468, 256, 256, 1) in 25.410072907805443 seconds.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/app/software/DeepCell/0.11.1-foss-2021b-CUDA-11.4.1/lib/python3.9/site-packages/deepcell_toolbox/deep_watershed.py:179: FutureWarning: `selem` is a deprecated argument name for `h_maxima`. It will be removed in version 1.0.Please use `footprint` instead.\n",
      "  markers = h_maxima(image=maxima,\n"
     ]
    }
   ],
   "source": [
    "# make predictions on testing data\n",
    "\n",
    "start = default_timer()\n",
    "test_images = prediction_model.predict(X_test)\n",
    "watershed_time = default_timer() - start\n",
    "\n",
    "print('Watershed segmentation of shape', test_images[0].shape,\n",
    "      'in', watershed_time, 'seconds.')\n",
    "\n",
    "masks = deep_watershed(\n",
    "    test_images,\n",
    "    min_distance=10,\n",
    "    detection_threshold=0.1,\n",
    "    distance_threshold=0.01,\n",
    "    exclude_border=False,\n",
    "    small_objects_threshold=0)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(468, 256, 256, 1)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 1,  1,  1, ...,  0,  0,  0],\n",
       "       [ 1,  1,  1, ...,  0,  0,  0],\n",
       "       [ 1,  1,  1, ...,  0,  0,  0],\n",
       "       ...,\n",
       "       [12, 12, 12, ...,  0,  0,  0],\n",
       "       [12, 12, 12, ...,  0,  0,  0],\n",
       "       [12, 12, 12, ...,  0,  0,  0]], dtype=uint8)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test[1,:,:,0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "0.95"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "io.imshow(X_test[2,:,:,0])\n",
    "plt.show()\n",
    "io.imshow(y_test[2,:,:,0])\n",
    "plt.show()\n",
    "syotil.csi(y_test[2,:,:,0], masks[2,:,:,0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(256, 256, 3)\n",
      "(256, 256)\n",
      "(1, 256, 256, 1)\n",
      "(1, 256, 256, 1)\n"
     ]
    }
   ],
   "source": [
    "# read CD3 data\n",
    "\n",
    "img=io.imread(\"images/square_patches/M872956_JML_Position8_CD3_img_patch256x256.png\")\n",
    "print(img.shape)\n",
    "X_cd3=img[:,:,2]\n",
    "#io.imshow(im)\n",
    "#plt.show()\n",
    "X_cd3=np.expand_dims(X_cd3, -1) \n",
    "X_cd3=np.expand_dims(X_cd3, 0) \n",
    "\n",
    "y_cd3=io.imread(\"images/square_patches/M872956_JML_Position8_CD3_masks_patch256x256.png\")\n",
    "print(y_cd3.shape)\n",
    "#io.imshow(mask_true)\n",
    "#plt.show()\n",
    "y_cd3=np.expand_dims(y_cd3, -1) \n",
    "y_cd3=np.expand_dims(y_cd3, 0) \n",
    "\n",
    "batch_size_cd3 = 1\n",
    "\n",
    "print(X_cd3.shape)\n",
    "print(y_cd3.shape)\n",
    "\n",
    "train_cd3 = datagen.flow(\n",
    "    {'X': X_cd3, 'y': y_cd3},\n",
    "    seed=seed,\n",
    "    transforms=transforms,\n",
    "    transforms_kwargs=transforms_kwargs,\n",
    "    min_objects=min_objects,\n",
    "    batch_size=batch_size)\n",
    "\n",
    "val_cd3 = datagen_val.flow(\n",
    "    {'X': X_cd3, 'y': y_cd3},\n",
    "    seed=seed,\n",
    "    transforms=transforms,\n",
    "    transforms_kwargs=transforms_kwargs,\n",
    "    min_objects=min_objects,\n",
    "    batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# make predictions on CD3 data\n",
    "start = default_timer()\n",
    "test_images = prediction_model.predict(X_cd3)\n",
    "watershed_time = default_timer() - start\n",
    "\n",
    "pred_cd3 = deep_watershed(\n",
    "    test_images,\n",
    "    min_distance=10,\n",
    "    detection_threshold=0.001,\n",
    "    distance_threshold=0.0,\n",
    "    exclude_border=False,\n",
    "    small_objects_threshold=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 256, 256, 1) (1, 256, 256, 1)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/app/software/scikit-image/0.19.1-foss-2021b/lib/python3.9/site-packages/skimage/io/_plugins/matplotlib_plugin.py:150: UserWarning: Low image data range; displaying image with stretched contrast.\n",
      "  lo, hi, cmap = _get_display_range(image)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "0.030303030303030304"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(y_cd3.shape, pred_cd3.shape)\n",
    "io.imshow(y_cd3[0,:,:,0])\n",
    "plt.show()\n",
    "io.imshow(pred_cd3[0,:,:,0])\n",
    "plt.show()\n",
    "syotil.csi(y_cd3[0,:,:,0], pred_cd3[0,:,:,0])\n",
    "# the cyto_fluo training data cells are bigger than CD3 cells"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('semantic_0', (None, 256, 256, 1)), ('semantic_1', (None, 256, 256, 1))]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from deepcell.applications import NuclearSegmentation\n",
    "app = NuclearSegmentation(prediction_model)\n",
    "[(layer.name, layer.output_shape) for layer in filter(lambda x: x.name.startswith('semantic_'), app.model.layers)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 256, 256, 1)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "0.10909090909090909"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y, tile_info = app._tile_input(X_cd3)\n",
    "print(y.shape)\n",
    "#print(tile_info)\n",
    "pred = app.predict(y, image_mpp=3)  \n",
    "prd = app._untile_output(pred, tile_info)\n",
    "io.imshow(prd[0,:,:,0])\n",
    "plt.show()\n",
    "syotil.csi(y_cd3[0,:,:,0], prd[0,:,:,0]) \n",
    "# default image_mpp 0.65, 0.03\n",
    "# image_mpp=1 0.111\n",
    "# image_mpp=2 0.212\n",
    "# image_mpp=3 0.109"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}