ForwardpassTests_Turbulence.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import torch\n",
    "import numpy as np\n",
    "import torch\n",
    "import math\n",
    "from torch import nn\n",
    "from utils import spectrum1d, FourierFilter, get_lr, DatasetFluid2D, train_model_batch\n",
    "from relaxed_gconv import TranSymDisNet\n",
    "from relaxed_gconv import IsoSymDisNet2D, RelaxedRotCNN2d, RelaxedRotGroupConv2d\n",
    "import matplotlib.pyplot as plt\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Isotropy Breaking"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/global/homes/r/rwang2/.conda/envs/e3nn/lib/python3.9/site-packages/scipy/signal/_spectral_py.py:2017: UserWarning: nperseg = 256 is greater than input length  = 128, using nperseg = 128\n",
      "  warnings.warn('nperseg = {0:d} is greater than input length '\n"
     ]
    }
   ],
   "source": [
    "sample = np.load(\"data/fluid/Re_180_1.npy\")\n",
    "sample = sample[None, :2]\n",
    "T = sample.shape[2]\n",
    "H = sample.shape[-2]\n",
    "W = sample.shape[-1]\n",
    "wave_numbers, spec_target = spectrum1d(sample)\n",
    "# torch_sample = torch.from_numpy(sample).reshape(1, T*2, H, W)\n",
    "# limits = torch.fft.rfftfreq(min(H, W))\n",
    "# separate_scales = []\n",
    "# for i in range(len(limits)-1):  \n",
    "#     separate_scale = FourierFilter(torch_sample, limits[i], limits[i+1]).reshape(2, T, H, W).real.data.numpy()\n",
    "#     np.save(\"data/fluid/scale_\" + str(i) + \".npy\", separate_scale)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "class RelaxedRotCNN2d(torch.nn.Module):\n",
    "    \"\"\"A small relaxed rotation 2d CNN model\"\"\"\n",
    "    def __init__(self, \n",
    "                 in_channels, \n",
    "                 out_channels, \n",
    "                 kernel_size, \n",
    "                 hidden_dim, \n",
    "                 group_order, # the order of 2d finite rotation group\n",
    "                 num_gconvs, # number of group conv layers\n",
    "                 num_filter_banks\n",
    "                ):\n",
    "        super(RelaxedRotCNN2d, self).__init__()\n",
    "        \n",
    "        self.group_order = group_order\n",
    "        self.gconvs = []\n",
    "        \n",
    "        theta = torch.tensor(2*np.pi/self.group_order).float()\n",
    "        self.lift_weights = torch.from_numpy(np.array([[np.cos(theta*i), np.sin(theta*i)] \n",
    "                                                       for i in range(self.group_order)])).float().to(device)\n",
    "\n",
    "        self.gconvs.append(RelaxedRotGroupConv2d(in_channels = in_channels,\n",
    "                                                 out_channels = hidden_dim,\n",
    "                                                 kernel_size = kernel_size,\n",
    "                                                 group_order = group_order,\n",
    "                                                 num_filter_banks = num_filter_banks,\n",
    "                                                 activation = True))\n",
    "        \n",
    "        for i in range(num_gconvs-2):\n",
    "            self.gconvs.append(RelaxedRotGroupConv2d(in_channels = hidden_dim,\n",
    "                                                     out_channels = hidden_dim,\n",
    "                                                     kernel_size = kernel_size,\n",
    "                                                     group_order = group_order,\n",
    "                                                     num_filter_banks = num_filter_banks,\n",
    "                                                     activation = True))\n",
    "            \n",
    "        self.gconvs.append(RelaxedRotGroupConv2d(in_channels = hidden_dim,\n",
    "                                                 out_channels = out_channels,\n",
    "                                                 kernel_size = kernel_size,\n",
    "                                                 group_order = group_order,\n",
    "                                                 num_filter_banks = num_filter_banks,\n",
    "                                                 activation = False))\n",
    "\n",
    "        self.gconvs = torch.nn.Sequential(*self.gconvs)\n",
    "\n",
    "    def forward(self, x, mean = True): \n",
    "        # average over h axis or not   \n",
    "        lift_x = torch.einsum(\"bvhw, gv -> bghw\", x, self.lift_weights).unsqueeze(1)\n",
    "        # print(lift_x.shape)\n",
    "        out = self.gconvs(lift_x)\n",
    "        if mean:\n",
    "            out = out.mean(2)\n",
    "        return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.13517518\n",
      "0.0549782\n",
      "0.04322963\n",
      "0.034796562\n",
      "0.0378226\n",
      "0.03774978\n",
      "0.032104418\n",
      "0.03496139\n",
      "0.04094966\n",
      "0.029606875\n",
      "0.0268868\n",
      "0.031922273\n",
      "0.028273333\n",
      "0.02398223\n",
      "0.024718586\n",
      "0.026186906\n",
      "0.03393533\n",
      "0.02510246\n",
      "0.021318126\n",
      "0.024922634\n",
      "0.016841982\n",
      "0.02036282\n",
      "0.023078306\n",
      "0.01963833\n",
      "0.016478788\n",
      "0.017017754\n",
      "0.017619789\n",
      "0.011292996\n",
      "0.009864543\n",
      "0.007892102\n",
      "0.007515028\n",
      "0.006741224\n",
      "0.0059868176\n",
      "0.0069780117\n",
      "0.00555392\n",
      "0.0048446637\n",
      "0.004468138\n",
      "0.0027797779\n",
      "0.003647254\n",
      "0.0039392477\n",
      "0.004310825\n",
      "0.0037047013\n",
      "0.003232369\n",
      "0.0033432103\n",
      "0.0035233533\n",
      "0.0033868582\n",
      "0.0038634418\n",
      "0.0031390742\n",
      "0.002901154\n",
      "0.0029668098\n",
      "0.002652898\n",
      "0.0031397797\n",
      "0.0026324373\n",
      "0.003302332\n",
      "0.0023944366\n",
      "0.0023526908\n",
      "0.002876326\n",
      "0.002263374\n",
      "0.0024724312\n",
      "0.0022981835\n",
      "0.0031814466\n",
      "0.0015934529\n",
      "0.0020082027\n",
      "0.0023205236\n",
      "0.28474233\n",
      "0.05045408\n",
      "0.034711577\n",
      "0.03924585\n",
      "0.029843573\n",
      "0.028858533\n",
      "0.025491463\n",
      "0.03121431\n",
      "0.034076843\n",
      "0.027059086\n",
      "0.03559157\n",
      "0.03276373\n",
      "0.03845707\n",
      "0.029391142\n",
      "0.022154856\n",
      "0.02280499\n",
      "0.027254613\n",
      "0.024309171\n",
      "0.028144494\n",
      "0.0194371\n",
      "0.020745749\n",
      "0.019186342\n",
      "0.020918483\n",
      "0.018696232\n",
      "0.015642809\n",
      "0.016662436\n",
      "0.01418441\n",
      "0.010487822\n",
      "0.010047579\n",
      "0.0076518613\n",
      "0.0068170074\n",
      "0.008408811\n",
      "0.004953498\n",
      "0.006442162\n",
      "0.005051601\n",
      "0.0045278785\n",
      "0.003711806\n",
      "0.0042431625\n",
      "0.005876998\n",
      "0.0034220866\n",
      "0.002638684\n",
      "0.003022389\n",
      "0.0040761363\n",
      "0.0037519615\n",
      "0.0031123506\n",
      "0.0023715568\n",
      "0.0023844428\n",
      "0.0036518585\n",
      "0.003971747\n",
      "0.003146605\n",
      "0.0028652896\n",
      "0.003075176\n",
      "0.0030679312\n",
      "0.0024741872\n",
      "0.0022712692\n",
      "0.0025936519\n",
      "0.0020403054\n",
      "0.002298717\n",
      "0.0025815752\n",
      "0.0019363934\n",
      "0.0018823199\n",
      "0.0015178241\n",
      "0.0015638074\n",
      "0.0018601018\n",
      "0.17878161\n",
      "0.059976637\n",
      "0.03488786\n",
      "0.03956812\n",
      "0.03635737\n",
      "0.036002394\n",
      "0.03472233\n",
      "0.02467295\n",
      "0.034448225\n",
      "0.033163685\n",
      "0.030936407\n",
      "0.027109012\n",
      "0.029968796\n",
      "0.025485085\n",
      "0.025817115\n",
      "0.025826985\n",
      "0.027269756\n",
      "0.028364036\n",
      "0.023076909\n",
      "0.022694325\n",
      "0.01580129\n",
      "0.02421079\n",
      "0.022551883\n",
      "0.016403958\n",
      "0.01626331\n",
      "0.010035465\n",
      "0.015356215\n",
      "0.011932645\n",
      "0.009192834\n",
      "0.007247585\n",
      "0.008885206\n",
      "0.0055968417\n",
      "0.005942769\n",
      "0.006220614\n",
      "0.0043442673\n",
      "0.004551945\n",
      "0.0047138715\n",
      "0.0055044703\n",
      "0.005080596\n",
      "0.004009565\n",
      "0.0039199484\n",
      "0.0037008133\n",
      "0.0034587632\n",
      "0.0032192767\n",
      "0.003727149\n",
      "0.0028227777\n",
      "0.0025852486\n",
      "0.003027238\n",
      "0.0034585053\n",
      "0.0029079819\n",
      "0.0026310857\n",
      "0.0026376122\n",
      "0.0027449715\n",
      "0.00287587\n",
      "0.0026036808\n",
      "0.0025294973\n",
      "0.0025894863\n",
      "0.0024268331\n",
      "0.0025762606\n",
      "0.0026586934\n",
      "0.0019042159\n",
      "0.0021728522\n",
      "0.0016786284\n",
      "0.0028307894\n",
      "0.19162115\n",
      "0.066066645\n",
      "0.042434912\n",
      "0.035912085\n",
      "0.021031758\n",
      "0.028838815\n",
      "0.033996757\n",
      "0.035087056\n",
      "0.030657766\n",
      "0.032605324\n",
      "0.021582002\n",
      "0.031425457\n",
      "0.030717684\n",
      "0.031593874\n",
      "0.023225104\n",
      "0.032826472\n",
      "0.037495527\n",
      "0.038981564\n",
      "0.024443246\n",
      "0.015842956\n",
      "0.019498851\n",
      "0.02505544\n",
      "0.01883196\n",
      "0.021707403\n",
      "0.015631424\n",
      "0.014760857\n",
      "0.013966515\n",
      "0.009465659\n",
      "0.006677044\n",
      "0.006937631\n",
      "0.00846308\n",
      "0.006950486\n",
      "0.0047586896\n",
      "0.00538869\n",
      "0.00483386\n",
      "0.005988491\n",
      "0.0047846246\n",
      "0.0036033466\n",
      "0.0033700443\n",
      "0.0040528933\n",
      "0.00290677\n",
      "0.0034780942\n",
      "0.0045766598\n",
      "0.0037926892\n",
      "0.003432356\n",
      "0.003095717\n",
      "0.0026449068\n",
      "0.0031381524\n",
      "0.0040568775\n",
      "0.0025709942\n",
      "0.002178579\n",
      "0.0026909069\n",
      "0.0033619478\n",
      "0.003350336\n",
      "0.0034432253\n",
      "0.0026998315\n",
      "0.0031621305\n",
      "0.0025590654\n",
      "0.0029790066\n",
      "0.0022568037\n",
      "0.0019521686\n",
      "0.0016325956\n",
      "0.00080482825\n",
      "0.002734594\n",
      "0.055060897\n",
      "0.05686926\n",
      "0.041489244\n",
      "0.049679022\n",
      "0.038895037\n",
      "0.045537133\n",
      "0.02153425\n",
      "0.03391274\n",
      "0.0224488\n",
      "0.027500931\n",
      "0.03843025\n",
      "0.029718691\n",
      "0.031144781\n",
      "0.026279608\n",
      "0.03200082\n",
      "0.02715842\n",
      "0.03539104\n",
      "0.026462628\n",
      "0.024261085\n",
      "0.023506666\n",
      "0.01974147\n",
      "0.021974102\n",
      "0.017384352\n",
      "0.023997398\n",
      "0.018467976\n",
      "0.016681135\n",
      "0.015550754\n",
      "0.011257087\n",
      "0.008940793\n",
      "0.006665117\n",
      "0.009633978\n",
      "0.007914912\n",
      "0.0066409204\n",
      "0.0056999414\n",
      "0.005416437\n",
      "0.0053283693\n",
      "0.004855328\n",
      "0.0039421194\n",
      "0.0029837638\n",
      "0.0042074267\n",
      "0.004033844\n",
      "0.0031912744\n",
      "0.003658354\n",
      "0.004207804\n",
      "0.0030489254\n",
      "0.0025150273\n",
      "0.0033419987\n",
      "0.0036603701\n",
      "0.0032146447\n",
      "0.0033860658\n",
      "0.0036252341\n",
      "0.0021437693\n",
      "0.0026323174\n",
      "0.0032516315\n",
      "0.0024306802\n",
      "0.0023137296\n",
      "0.0028944325\n",
      "0.001852097\n",
      "0.0029020812\n",
      "0.0028449865\n",
      "0.0014429436\n",
      "0.0020830103\n",
      "0.0009709282\n",
      "0.0024548701\n"
     ]
    }
   ],
   "source": [
    "all_equiv_errors = []\n",
    "# Average over five runs with different random seeds\n",
    "for seed in range(5):\n",
    "    torch.manual_seed(seed)\n",
    "    torch.cuda.manual_seed_all(seed)\n",
    "    equiv_errors = []\n",
    "    for s in range(64):\n",
    "        in_channels = 1\n",
    "        out_channels = 1\n",
    "        kernel_size = 5\n",
    "        hidden_dim = 16\n",
    "        num_filter_basis = 1\n",
    "        batch_size = 2\n",
    "        train_indices = np.arange(0, 1800)\n",
    "\n",
    "        data_direc = \"data/fluid/scale_\" + str(s) + \".npy\"\n",
    "        train_set = DatasetFluid2D(direc = data_direc, sample_list = train_indices)\n",
    "        train_loader = torch.utils.data.DataLoader(train_set, batch_size = batch_size, shuffle = True, num_workers = 2)\n",
    "\n",
    "        model = RelaxedRotCNN2d(in_channels = in_channels, \n",
    "                                out_channels = out_channels, \n",
    "                                kernel_size = kernel_size, \n",
    "                                hidden_dim = hidden_dim, \n",
    "                                group_order = 4, # the order of 2d finite rotation group\n",
    "                                num_gconvs = 3, # number of group conv layers\n",
    "                                num_filter_banks = 1).to(device)\n",
    "        with torch.no_grad():\n",
    "            all_outs = []\n",
    "            for xx, yy in train_loader:\n",
    "                xx = xx.to(device)\n",
    "                yy = yy.to(device)\n",
    "                preds = model(xx.to(device), mean = False)\n",
    "                all_outs.append(preds)\n",
    "            all_outs = torch.cat(all_outs, dim = 0) \n",
    "        # Normalization\n",
    "        all_outs = (all_outs - torch.min(all_outs))/(torch.max(all_outs) - torch.min(all_outs))\n",
    "        # Equiv Error\n",
    "        equiv_errors.append(torch.mean(torch.abs(all_outs - all_outs[:,:,:1])).cpu().data.numpy())\n",
    "        print(equiv_errors[-1])\n",
    "    all_equiv_errors.append(equiv_errors)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "all_equiv_errors = np.array(all_equiv_errors)\n",
    "std_equiv_error = np.std(all_equiv_errors, axis = 0)\n",
    "avg_equiv_error = np.mean(all_equiv_errors, axis = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x450 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig=plt.figure(figsize=(6, 4.5))\n",
    "plt.plot(wave_numbers[1:], avg_equiv_error, color = \"royalblue\", linewidth = 3)\n",
    "plt.fill_between(wave_numbers[1:], (avg_equiv_error-std_equiv_error), (avg_equiv_error+std_equiv_error), alpha=.3)\n",
    "plt.ylabel(\"Equivariance Error\", size = 15)\n",
    "plt.xlabel(\"Wave Number\", size = 14.5)\n",
    "plt.title(\"Rotation Equiv Errors vs. WaveNumbers (GroupConv Forward Passing)\", size = 17)\n",
    "plt.grid(axis='both', alpha=.3)\n",
    "plt.xscale(\"log\")\n",
    "\n",
    "# Remove borders\n",
    "plt.gca().spines[\"top\"].set_alpha(0.0)    \n",
    "plt.gca().spines[\"bottom\"].set_alpha(0.3)\n",
    "plt.gca().spines[\"right\"].set_alpha(0.0)    \n",
    "plt.gca().spines[\"left\"].set_alpha(0.3)   \n",
    "# plt.show()\n",
    "plt.savefig(\"iso_breaking_forward_passing.png\", dpi = 200, bbox_inches = 'tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Homogeneity Breaking"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "in_channels = 2\n",
    "out_channels = 1\n",
    "kernel_size = 3\n",
    "h_size = 256\n",
    "w_size = 128\n",
    "hidden_dim = 16\n",
    "num_filter_basis = 1\n",
    "batch_size = 128\n",
    "learning_rate = 0.001\n",
    "decay_rate = 0.9\n",
    "num_epoch = 100\n",
    "train_indices = np.arange(0, 1800)\n",
    "\n",
    "all_equiv_errors_seeds = []\n",
    "for seed in range(5):\n",
    "    torch.manual_seed(seed)\n",
    "    torch.cuda.manual_seed_all(seed)\n",
    "    \n",
    "    data_direc = \"data/fluid/Re_180_1.npy\"\n",
    "\n",
    "    train_set = DatasetFluid2D(direc = data_direc, sample_list = train_indices)\n",
    "    train_loader = torch.utils.data.DataLoader(train_set, batch_size = batch_size, shuffle = True, num_workers = 2)\n",
    "\n",
    "    model = TranSymDisNet(in_channels = in_channels, \n",
    "                          out_channels = out_channels, \n",
    "                          kernel_size = kernel_size, \n",
    "                          h_size = h_size, \n",
    "                          w_size = w_size, \n",
    "                          hidden_dim = hidden_dim, \n",
    "                          num_filter_basis = num_filter_basis).to(device)\n",
    "\n",
    "    with torch.no_grad():\n",
    "        all_outs = []\n",
    "        for xx, yy in train_loader:\n",
    "            xx = xx.to(device)\n",
    "            yy = yy.to(device)\n",
    "            preds = model(xx.to(device))\n",
    "            all_outs.append(preds)\n",
    "        all_outs = torch.cat(all_outs, dim = 0) \n",
    "    all_outs = all_outs.cpu().data.numpy()\n",
    "    all_outs = (all_outs - all_outs.min((2), keepdims = True))/(all_outs.max((2), keepdims = True) - all_outs.min((2), keepdims = True))\n",
    "    \n",
    "#     # Calculate the equivariance error based on the learnt relaxed weights\n",
    "#     # \\sum_{g \\in G} |w(e) - w(g)|\n",
    "    equiv_error_y = np.mean(np.abs(all_outs - all_outs[:,:,63:64]), axis = -1).mean(0).mean(0)\n",
    "    # break\n",
    "    all_equiv_errors_seeds.append(equiv_error_y)\n",
    "\n",
    "# std_equiv_error = np.std(all_equiv_errors_seeds, axis = 0)\n",
    "# avg_equiv_error = np.mean(all_equiv_errors_seeds, axis = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "std_equiv_error = np.std(all_equiv_errors_seeds, axis = 0)\n",
    "avg_equiv_error = np.mean(all_equiv_errors_seeds, axis = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig=plt.figure(figsize=(5, 4))\n",
    "# remove boundary effects\n",
    "# avg_equiv_error = avg_equiv_error[3:-3]\n",
    "# std_equiv_error = std_equiv_error[3:-3]\n",
    "plt.plot(avg_equiv_error,  linewidth = 3.5,  alpha=.8)\n",
    "plt.fill_between(np.arange(len(avg_equiv_error)), (avg_equiv_error-std_equiv_error), (avg_equiv_error+std_equiv_error), alpha=.5)\n",
    "plt.ylabel(\"Equivariance Error\", size = 15)\n",
    "plt.xlabel(\"Y Axis\", size = 14)\n",
    "plt.title(\"Translation Equiv Error along Y Axis (GroupConv Forward Passing)\", size = 16)\n",
    "plt.gca().spines[\"top\"].set_alpha(0.0)    \n",
    "plt.gca().spines[\"bottom\"].set_alpha(1)\n",
    "plt.gca().spines[\"right\"].set_alpha(0.0)    \n",
    "plt.gca().spines[\"left\"].set_alpha(1)   \n",
    "plt.grid(True, linestyle='-.')\n",
    "# plt.show()\n",
    "plt.savefig(\"translation_equiv_forward_passing.png\", dpi = 200, bbox_inches = 'tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "e3nn",
   "language": "python",
   "name": "e3nn"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.17"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}