neural_operator_practise.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training a TFNO on Darcy-Flow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import matplotlib.pyplot as plt\n",
    "import sys\n",
    "from neuralop.models import TFNO\n",
    "from neuralop import Trainer\n",
    "from neuralop.datasets import load_darcy_flow_small, burgers\n",
    "from neuralop.utils import count_model_params\n",
    "from neuralop import LpLoss, H1Loss\n",
    "\n",
    "device = 'cuda'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading test db at resolution 32 with 50 samples and batch-size=32\n"
     ]
    }
   ],
   "source": [
    "train_loader, test_loaders, data_processor = load_darcy_flow_small(\n",
    "        n_train=1000, batch_size=32,\n",
    "        test_resolutions=[16, 32], n_tests=[100, 50],\n",
    "        test_batch_sizes=[32, 32],\n",
    "        positional_encoding=True\n",
    ")\n",
    "data_processor = data_processor.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([16, 16])"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x[0].size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{16: <torch.utils.data.dataloader.DataLoader at 0x7fe47c6626d0>,\n",
       " 32: <torch.utils.data.dataloader.DataLoader at 0x7fe47c510f70>}"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_loaders"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 700x700 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_samples = test_loaders[16].dataset\n",
    "\n",
    "fig = plt.figure(figsize=(7, 7))\n",
    "for index in range(3):\n",
    "    data = test_samples[index]\n",
    "    data = data_processor.preprocess(data, batched=False)\n",
    "    # Input x\n",
    "    x = data['x']\n",
    "    # Ground-truth\n",
    "    y = data['y']\n",
    "    # Model prediction\n",
    "    #out = model(x.unsqueeze(0))\n",
    "\n",
    "    ax = fig.add_subplot(3, 3, index*3 + 1)\n",
    "    ax.imshow(x[0].cpu(), cmap='gray')\n",
    "    if index == 0:\n",
    "        ax.set_title('Input x')\n",
    "    plt.xticks([], [])\n",
    "    plt.yticks([], [])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## We create a tensorized FNO model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Our model has 523257 parameters.\n"
     ]
    }
   ],
   "source": [
    "model = TFNO(n_modes=(16, 16), hidden_channels=32, projection_channels=64, factorization='tucker', rank=0.42)\n",
    "model = model.to(device)\n",
    "\n",
    "n_params = count_model_params(model)\n",
    "print(f'\\nOur model has {n_params} parameters.')\n",
    "sys.stdout.flush()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create the optimizer\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "optimizer = torch.optim.Adam(model.parameters(),\n",
    "                                lr=8e-3,\n",
    "                                weight_decay=1e-4)\n",
    "scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(optimizer, T_max=30)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Creating the losses"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "l2loss = LpLoss(d=2, p=2)\n",
    "h1loss = H1Loss(d=2)\n",
    "\n",
    "train_loss = h1loss\n",
    "eval_losses={'h1': h1loss, 'l2': l2loss}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "### MODEL ###\n",
      " TFNO(\n",
      "  (fno_blocks): FNOBlocks(\n",
      "    (convs): SpectralConv(\n",
      "      (weight): ModuleList(\n",
      "        (0-3): 4 x ComplexTuckerTensor(shape=(32, 32, 16, 9), rank=(26, 26, 13, 7))\n",
      "      )\n",
      "    )\n",
      "    (fno_skips): ModuleList(\n",
      "      (0-3): 4 x Conv2d(32, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "    )\n",
      "  )\n",
      "  (lifting): MLP(\n",
      "    (fcs): ModuleList(\n",
      "      (0): Conv2d(3, 256, kernel_size=(1, 1), stride=(1, 1))\n",
      "      (1): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1))\n",
      "    )\n",
      "  )\n",
      "  (projection): MLP(\n",
      "    (fcs): ModuleList(\n",
      "      (0): Conv2d(32, 64, kernel_size=(1, 1), stride=(1, 1))\n",
      "      (1): Conv2d(64, 1, kernel_size=(1, 1), stride=(1, 1))\n",
      "    )\n",
      "  )\n",
      ")\n",
      "\n",
      "### OPTIMIZER ###\n",
      " Adam (\n",
      "Parameter Group 0\n",
      "    amsgrad: False\n",
      "    betas: (0.9, 0.999)\n",
      "    capturable: False\n",
      "    differentiable: False\n",
      "    eps: 1e-08\n",
      "    foreach: None\n",
      "    fused: None\n",
      "    initial_lr: 0.008\n",
      "    lr: 0.008\n",
      "    maximize: False\n",
      "    weight_decay: 0.0001\n",
      ")\n",
      "\n",
      "### SCHEDULER ###\n",
      " <torch.optim.lr_scheduler.CosineAnnealingLR object at 0x7fe5fdfc6be0>\n",
      "\n",
      "### LOSSES ###\n",
      "\n",
      " * Train: <neuralop.losses.data_losses.H1Loss object at 0x7fe5fe063e80>\n",
      "\n",
      " * Test: {'h1': <neuralop.losses.data_losses.H1Loss object at 0x7fe5fe063e80>, 'l2': <neuralop.losses.data_losses.LpLoss object at 0x7fe5fded0610>}\n"
     ]
    }
   ],
   "source": [
    "print('\\n### MODEL ###\\n', model)\n",
    "print('\\n### OPTIMIZER ###\\n', optimizer)\n",
    "print('\\n### SCHEDULER ###\\n', scheduler)\n",
    "print('\\n### LOSSES ###')\n",
    "print(f'\\n * Train: {train_loss}')\n",
    "print(f'\\n * Test: {eval_losses}')\n",
    "sys.stdout.flush()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create the trainer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "self.override_load_to_device=False\n",
      "self.overrides_loss=False\n"
     ]
    }
   ],
   "source": [
    "trainer = Trainer(model=model, n_epochs=20,\n",
    "                  device=device,\n",
    "                  data_processor=data_processor,\n",
    "                  wandb_log=False,\n",
    "                  log_test_interval=3,\n",
    "                  use_distributed=False,\n",
    "                  verbose=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Actually train the model on our small Darcy-Flow dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'32_h1': 0.3513032817840576, '32_l2': 0.21183603763580322}"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainer.train(train_loader=train_loader,\n",
    "              test_loaders=test_loaders,\n",
    "              optimizer=optimizer,\n",
    "              scheduler=scheduler,\n",
    "              regularizer=False,\n",
    "              training_loss=train_loss,\n",
    "              eval_losses=eval_losses)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 700x700 with 9 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_samples = test_loaders[32].dataset\n",
    "\n",
    "fig = plt.figure(figsize=(7, 7))\n",
    "for index in range(3):\n",
    "    data = test_samples[index]\n",
    "    data = data_processor.preprocess(data, batched=False)\n",
    "    # Input x\n",
    "    x = data['x']\n",
    "    # Ground-truth\n",
    "    y = data['y']\n",
    "    # Model prediction\n",
    "    out = model(x.unsqueeze(0))\n",
    "\n",
    "    ax = fig.add_subplot(3, 3, index*3 + 1)\n",
    "    ax.imshow(x[0].cpu(), cmap='gray')\n",
    "    if index == 0:\n",
    "        ax.set_title('Input x')\n",
    "    plt.xticks([], [])\n",
    "    plt.yticks([], [])\n",
    "\n",
    "    ax = fig.add_subplot(3, 3, index*3 + 2)\n",
    "    ax.imshow(y.squeeze().cpu())\n",
    "    if index == 0:\n",
    "        ax.set_title('Ground-truth y')\n",
    "    plt.xticks([], [])\n",
    "    plt.yticks([], [])\n",
    "\n",
    "    ax = fig.add_subplot(3, 3, index*3 + 3)\n",
    "    ax.imshow(out.squeeze().detach().cpu().numpy())\n",
    "    if index == 0:\n",
    "        ax.set_title('Model prediction')\n",
    "    plt.xticks([], [])\n",
    "    plt.yticks([], [])\n",
    "\n",
    "fig.suptitle('Inputs, ground-truth output and prediction.', y=0.98)\n",
    "plt.tight_layout()\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 32, 32])"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y[0].size()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "NeuralOperator",
   "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": "3.8.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}