03_train_seq2seq.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from fastai.vision.all import *\n",
    "from moving_mnist.models.seq2seq import *\n",
    "from moving_mnist.models.conv_rnn import StackUnstack, StackLoss, MultiImageDice\n",
    "from moving_mnist.data import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Quadro RTX 8000\n"
     ]
    }
   ],
   "source": [
    "if torch.cuda.is_available():\n",
    "    torch.cuda.set_device(0)\n",
    "    print(torch.cuda.get_device_name())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train Example:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We wil predict:\n",
    "- `n_in`: 5 images\n",
    "- `n_out`: 5 images  \n",
    "- `n_obj`: up to 3 objects"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Path('/home/tcapelle/SteadySun/moving_mnist')"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Path.cwd()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "DATA_PATH = Path.cwd()/'data'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "class ImageStack(Transform):\n",
    "    order=89\n",
    "    def __init__(self, dim=1): store_attr()\n",
    "    \n",
    "    def encodes(self, x:ImageSeq): return torch.stack(x, dim=self.dim)\n",
    "    def decodes(self, x:ImageSeq): return torch.unbind(x, dim=self.dim)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "ds = MovingMNIST(DATA_PATH, n_in=5, n_out=5, n_obj=[1,2], th=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_tl = TfmdLists(range(12500), ImageTupleTransform(ds))\n",
    "valid_tl = TfmdLists(range(100), ImageTupleTransform(ds))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "dls = DataLoaders.from_dsets(train_tl, valid_tl, bs=64,\n",
    "                             after_batch=[Normalize.from_stats(*mnist_stats)]).cuda()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "mse_loss = StackLoss(MSELossFlat(axis=1))\n",
    "metrics = []"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Left: Input, Right: Target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x777.6 with 18 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "dls.show_batch()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "b = dls.one_batch()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{tuple: [{moving_mnist.data.ImageSeq: [fastai.torch_core.TensorImageBW,\n",
       "    fastai.torch_core.TensorImageBW,\n",
       "    fastai.torch_core.TensorImageBW,\n",
       "    fastai.torch_core.TensorImageBW,\n",
       "    fastai.torch_core.TensorImageBW]},\n",
       "  {moving_mnist.data.ImageSeq: [fastai.torch_core.TensorImageBW,\n",
       "    fastai.torch_core.TensorImageBW,\n",
       "    fastai.torch_core.TensorImageBW,\n",
       "    fastai.torch_core.TensorImageBW,\n",
       "    fastai.torch_core.TensorImageBW]}]}"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "explode_types(b)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([64, 1, 64, 64])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x,y = b \n",
    "x[0].shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": "true"
   },
   "source": [
    "## Baseline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`StackUnstack` takes cares of stacking the list of images into a fat tensor, and unstacking them at the end, we will need to modify our loss function to take a list of tensors as input and target."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = StackUnstack(Seq2Seq(5, ch_out=1), dim=1).cuda()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As the `ImageSeq` is a `tuple` of images, we will need to stack them to compute loss."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn = Learner(dls, model, loss_func=mse_loss, metrics=metrics, cbs=[EncoderLoss(mse_loss)])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I have a weird bug that if I use `nn.LeakyReLU` after doing `learn.lr_find()` the model does not train (the loss get stucked)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "SuggestedLRs(lr_min=0.03019951581954956, lr_steep=0.0002290867705596611)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "learn.lr_find()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>0.480382</td>\n",
       "      <td>0.371032</td>\n",
       "      <td>00:28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>0.343120</td>\n",
       "      <td>0.329683</td>\n",
       "      <td>00:28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>0.314175</td>\n",
       "      <td>0.300663</td>\n",
       "      <td>00:28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>0.292787</td>\n",
       "      <td>0.257575</td>\n",
       "      <td>00:28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>0.276540</td>\n",
       "      <td>0.277418</td>\n",
       "      <td>00:28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>0.263274</td>\n",
       "      <td>0.274320</td>\n",
       "      <td>00:28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>0.254291</td>\n",
       "      <td>0.246722</td>\n",
       "      <td>00:28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>0.247428</td>\n",
       "      <td>0.268464</td>\n",
       "      <td>00:28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>0.241123</td>\n",
       "      <td>0.241932</td>\n",
       "      <td>00:28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>0.239736</td>\n",
       "      <td>0.229110</td>\n",
       "      <td>00:28</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.fit_one_cycle(10, 1e-3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "p,t = learn.get_preds()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As you can see, the results is a list of 5 tensors with 100 samples each."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5, torch.Size([100, 1, 64, 64]))"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(p), p[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_res(t, idx, argmax=False):\n",
    "    if argmax:\n",
    "        im_seq = ImageSeq.create([t[i][idx].argmax(0).unsqueeze(0) for i in range(5)], TensorMask)\n",
    "    else:\n",
    "        im_seq = ImageSeq.create([t[i][idx] for i in range(5)])\n",
    "    im_seq.show(figsize=(8,4));"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "k = random.randint(0,99)\n",
    "show_res(t,k)\n",
    "show_res(p,k, False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Teacher Forcing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = StackUnstack(Seq2Seq(5, ch_out=1, ks=3, n_layers=1, debug=False), dim=1).cuda()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn = Learner(dls, model, loss_func=mse_loss, metrics=metrics, \n",
    "                cbs=[TeacherForcing(4), EncoderLoss(mse_loss, alpha=1)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "SuggestedLRs(lr_min=0.06309573650360108, lr_steep=0.0004786300996784121)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "learn.lr_find()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>0.334137</td>\n",
       "      <td>0.298063</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>0.214372</td>\n",
       "      <td>0.225534</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>0.195803</td>\n",
       "      <td>0.208463</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>0.187567</td>\n",
       "      <td>0.193403</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>0.189288</td>\n",
       "      <td>0.180637</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>0.182708</td>\n",
       "      <td>0.166872</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>0.173564</td>\n",
       "      <td>0.170718</td>\n",
       "      <td>00:32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>0.169031</td>\n",
       "      <td>0.167578</td>\n",
       "      <td>00:32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>0.163887</td>\n",
       "      <td>0.168154</td>\n",
       "      <td>00:32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>0.161908</td>\n",
       "      <td>0.150254</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>10</td>\n",
       "      <td>0.159680</td>\n",
       "      <td>0.164322</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>11</td>\n",
       "      <td>0.159712</td>\n",
       "      <td>0.152653</td>\n",
       "      <td>00:33</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.fit_one_cycle(12, 1e-3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "p,t = learn.get_preds()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As you can see, the results is a list of 5 tensors with 100 samples each."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5, torch.Size([100, 1, 64, 64]))"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(p), p[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_res(t, idx, argmax=False):\n",
    "    if argmax:\n",
    "        im_seq = ImageSeq.create([t[i][idx].argmax(0).unsqueeze(0) for i in range(5)], TensorMask)\n",
    "    else:\n",
    "        im_seq = ImageSeq.create([t[i][idx] for i in range(5)])\n",
    "    im_seq.show(figsize=(8,4));"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "k = random.randint(0,99)\n",
    "show_res(t,k)\n",
    "show_res(p,k, False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## PhyCell"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = StackUnstack(Seq2SeqPhy(5, ch_out=1, ks=3, n_layers=3, debug=False), dim=1).cuda()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn = Learner(dls, model, loss_func=mse_loss, metrics=metrics, \n",
    "                cbs=[TeacherForcing(6), EncoderLoss(mse_loss, alpha=1)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "SuggestedLRs(lr_min=0.04365158379077912, lr_steep=9.120108734350652e-05)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "learn.lr_find()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>0.612359</td>\n",
       "      <td>0.535470</td>\n",
       "      <td>01:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>0.393701</td>\n",
       "      <td>0.537646</td>\n",
       "      <td>01:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>0.329480</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>01:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>0.287261</td>\n",
       "      <td>0.414644</td>\n",
       "      <td>01:07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>0.255289</td>\n",
       "      <td>0.373533</td>\n",
       "      <td>01:09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>0.245856</td>\n",
       "      <td>0.315296</td>\n",
       "      <td>01:09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>0.238928</td>\n",
       "      <td>0.297825</td>\n",
       "      <td>01:10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>0.228544</td>\n",
       "      <td>0.289100</td>\n",
       "      <td>01:10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>0.219233</td>\n",
       "      <td>0.304926</td>\n",
       "      <td>01:10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>0.213856</td>\n",
       "      <td>0.302157</td>\n",
       "      <td>01:10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>10</td>\n",
       "      <td>0.210935</td>\n",
       "      <td>0.309541</td>\n",
       "      <td>01:10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>11</td>\n",
       "      <td>0.208779</td>\n",
       "      <td>0.292696</td>\n",
       "      <td>01:10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.fit_one_cycle(12, 1e-3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "p,t = learn.get_preds()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As you can see, the results is a list of 5 tensors with 100 samples each."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5, torch.Size([100, 1, 64, 64]))"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(p), p[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_res(t, idx, argmax=False):\n",
    "    if argmax:\n",
    "        im_seq = ImageSeq.create([t[i][idx].argmax(0).unsqueeze(0) for i in range(5)], TensorMask)\n",
    "    else:\n",
    "        im_seq = ImageSeq.create([t[i][idx] for i in range(5)])\n",
    "    im_seq.show(figsize=(8,4));"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "k = random.randint(0,99)\n",
    "show_res(t,k)\n",
    "show_res(p,k, False)"
   ]
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}