two_objects_model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Two objects dataset\n",
    "Simple version of two objects dataset. Added \"near\" relation. No \"on\" relation. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 309,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "import sampling\n",
    "from evaluation import evaluate_sample\n",
    "from environment import grab_frame, reset_env, create_shape\n",
    "from predicates import AtomPredicate, AtomRelation, AtomColor, AtomObject\n",
    "\n",
    "frame_size = (64, 64)\n",
    "dataset_path = 'data/two_objects_dataset_64'\n",
    "\n",
    "def sample_func():\n",
    "    bounds = (4.0, 4.0, 0.0)\n",
    "    obj_1 = sampling.ObjectOnTable.sample(origin=(0, 0, 1.0), max_bounds=bounds, max_rotation=2 * np.pi, size=2.0)\n",
    "    obj_2 = sampling.ObjectOnTable.sample_no_intersection(\n",
    "        origin=(0, 0, 1.0), max_bounds=bounds, max_rotation=2 * np.pi, size=2.0, existing_objects=[obj_1], min_distance=2.0\n",
    "    )    \n",
    "    \n",
    "    predicates = (\n",
    "        sampling.get_on_table_relation(obj_1, bounds, add_positional=False) + \n",
    "        sampling.get_on_table_relation(obj_2, bounds, add_positional=False)\n",
    "        #sampling.get_near_relation([obj_1, obj_2])\n",
    "    )\n",
    "\n",
    "    if obj_1.obj_type != obj_2.obj_type and obj_1.color != obj_2.color:\n",
    "        predicates.extend([\n",
    "            (AtomPredicate(AtomRelation.ON, obj_1.obj_type, obj_2.color, AtomObject.TABLE, AtomColor.NO_COLOR), 0.0),\n",
    "            (AtomPredicate(AtomRelation.ON, obj_2.obj_type, obj_1.color, AtomObject.TABLE, AtomColor.NO_COLOR), 0.0),\n",
    "        ])    \n",
    "        predicates.extend(sampling.get_false_predicates(predicates, 2))\n",
    "    else:\n",
    "        predicates.extend(sampling.get_false_predicates(predicates, 4))\n",
    "\n",
    "    reset_env()\n",
    "    create_shape(obj_1.obj_type, obj_1.color, obj_1.position, obj_1.orientation, obj_1.size)    \n",
    "    create_shape(obj_2.obj_type, obj_2.color, obj_2.position, obj_2.orientation, obj_2.size)        \n",
    "    frame = grab_frame(cam_pos=(0, -10, 12), cam_target=(0, -1, 0), light_dir=(-6, 1, 10), frame_size=frame_size)\n",
    "    return frame, predicates\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 310,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(blue cube on table) 1.0\n",
      "(purple pyramid on table) 1.0\n",
      "(purple cube on table) 0.0\n",
      "(blue pyramid on table) 0.0\n",
      "(purple sphere on table) 0.0\n",
      "(green pyramid on table) 0.0\n"
     ]
    }
   ],
   "source": [
    "evaluate_sample(sample_func)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 312,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Generating 128 * 128 = 16384 samples\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 128/128 [07:15<00:00,  3.41s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Generating 8 * 128 = 1024 samples\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 8/8 [00:28<00:00,  3.57s/it]\n"
     ]
    }
   ],
   "source": [
    "# Run only once!\n",
    "from batching import create_dataset\n",
    "\n",
    "create_dataset(dataset_path, num_batches=128, batch_size=128, sample_func=sample_func)\n",
    "create_dataset(dataset_path + '_test', num_batches=8, batch_size=128, sample_func=sample_func)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 337,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_8\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_17 (InputLayer)           [(None, 64, 64, 3)]  0                                            \n",
      "__________________________________________________________________________________________________\n",
      "rescaling_8 (Rescaling)         (None, 64, 64, 3)    0           input_17[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_24 (Conv2D)              (None, 64, 64, 8)    224         rescaling_8[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_24 (MaxPooling2D) (None, 32, 32, 8)    0           conv2d_24[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_25 (Conv2D)              (None, 32, 32, 8)    584         max_pooling2d_24[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_25 (MaxPooling2D) (None, 16, 16, 8)    0           conv2d_25[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_26 (Conv2D)              (None, 16, 16, 16)   1168        max_pooling2d_25[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_26 (MaxPooling2D) (None, 8, 8, 16)     0           conv2d_26[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "flatten_8 (Flatten)             (None, 1024)         0           max_pooling2d_26[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "input_18 (InputLayer)           [(None, 33)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "dense_32 (Dense)                (None, 64)           65600       flatten_8[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "dense_33 (Dense)                (None, 64)           2176        input_18[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_8 (Concatenate)     (None, 128)          0           dense_32[0][0]                   \n",
      "                                                                 dense_33[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_34 (Dense)                (None, 128)          16512       concatenate_8[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "dense_35 (Dense)                (None, 1)            129         dense_34[0][0]                   \n",
      "==================================================================================================\n",
      "Total params: 86,393\n",
      "Trainable params: 86,393\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.keras import layers\n",
    "from tensorflow.keras.models import Model\n",
    "\n",
    "frame_inputs = layers.Input(shape=(*frame_size, 3))\n",
    "x = layers.Rescaling(1.0 / 255.0)(frame_inputs)\n",
    "\n",
    "for size in (8, 8, 16):\n",
    "    x = layers.Conv2D(size, 3, padding=\"same\", activation='relu')(x)\n",
    "    x = layers.MaxPooling2D(2, padding=\"same\")(x)        \n",
    "\n",
    "x = layers.Flatten()(x)\n",
    "conv_branch_outputs = layers.Dense(64, activation='relu')(x)\n",
    "\n",
    "predicate_inputs = layers.Input(shape=(33,))\n",
    "predicate_branch_outputs = layers.Dense(64, activation='relu')(predicate_inputs)\n",
    "\n",
    "x = layers.concatenate([conv_branch_outputs, predicate_branch_outputs], axis=-1)\n",
    "x = layers.Dense(128, activation='relu')(x)\n",
    "outputs = layers.Dense(1, activation='sigmoid')(x)\n",
    "\n",
    "model = Model([frame_inputs, predicate_inputs], outputs)\n",
    "model.summary()\n",
    "model.compile(optimizer=\"adam\", loss=\"binary_crossentropy\", metrics=[\"accuracy\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "metadata": {},
   "outputs": [],
   "source": [
    "from batching import load_as_tf_dataset\n",
    "training_set = load_as_tf_dataset(dataset_path).shuffle(128).batch(32)\n",
    "testing_set = load_as_tf_dataset(dataset_path + \"_test\").batch(32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 314,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Datasets contain 256 files in total.\n",
      "Datasets contain 16 files in total.\n"
     ]
    }
   ],
   "source": [
    "from batching import load_datasets\n",
    "\n",
    "training_set = load_datasets([\"data/one_object_dataset_64\", dataset_path]).shuffle(128).batch(32)\n",
    "testing_set = load_datasets([\"data/one_object_dataset_64_test\", dataset_path + \"_test\"]).batch(32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 338,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.load_weights(\"models/one_object_64input_8_8_16conv_64_128dense.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 369,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "7680/7680 [==============================] - 83s 11ms/step - loss: 0.0363 - accuracy: 0.9588 - val_loss: 0.0555 - val_accuracy: 0.9536\n",
      "Epoch 2/5\n",
      "7680/7680 [==============================] - 83s 11ms/step - loss: 0.0353 - accuracy: 0.9592 - val_loss: 0.0549 - val_accuracy: 0.9541\n",
      "Epoch 3/5\n",
      "7680/7680 [==============================] - 83s 11ms/step - loss: 0.0347 - accuracy: 0.9594 - val_loss: 0.0542 - val_accuracy: 0.9535\n",
      "Epoch 4/5\n",
      "7680/7680 [==============================] - 82s 11ms/step - loss: 0.0339 - accuracy: 0.9596 - val_loss: 0.0528 - val_accuracy: 0.9533\n",
      "Epoch 5/5\n",
      "7680/7680 [==============================] - 82s 11ms/step - loss: 0.0334 - accuracy: 0.9599 - val_loss: 0.0538 - val_accuracy: 0.9542\n",
      "CPU times: user 6min 39s, sys: 40.4 s, total: 7min 20s\n",
      "Wall time: 6min 53s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "model.optimizer.learning_rate = 1e-4\n",
    "progress = model.fit(training_set, validation_data=testing_set, epochs=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 370,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from utils import draw_progress\n",
    "draw_progress(progress, fit_line=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 373,
   "metadata": {},
   "outputs": [],
   "source": [
    "probes = sampling.get_on_table_probes(add_positional=True)\n",
    "#probes = sampling.get_near_probes()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(324, 33)"
      ]
     },
     "execution_count": 197,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "probes.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 380,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(purple sphere on table) 1.0\n",
      "(purple cube on table) 1.0\n",
      "(yellow pyramid on table) 0.0\n",
      "(green cube on table) 0.0\n",
      "(purple pyramid on table) 0.0\n",
      "(blue cube on table) 0.0\n",
      "---\n",
      "(purple cube on table) 0.99999535\n",
      "(purple cube on_near_side_of table) 0.9992373\n",
      "(purple sphere on table) 0.99876463\n",
      "(purple cube on_left_side_of table) 0.9765089\n"
     ]
    }
   ],
   "source": [
    "evaluate_sample(sample_func, model, probes)"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "d8c0228d123c1147fccbb2d1956029d0169d2e4b1fdd13bb4e68fe309dcbd4e2"
  },
  "kernelspec": {
   "display_name": "Python 3.9.7 64-bit ('ml': conda)",
   "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": "3.9.7"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}