train_model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "# import torchsummary\n",
    "import torchmetrics\n",
    "from IPython.display import clear_output\n",
    "from matplotlib import pyplot as plt\n",
    "from matplotlib.pyplot import figure\n",
    "import numpy as np\n",
    "plt.rcParams['text.usetex'] = True\n",
    "\n",
    "import os\n",
    "from h5 import *\n",
    "\n",
    "import time\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "device = 'cuda:0' if torch.cuda.is_available() else 'cpu'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from triqs.operators import *\n",
    "from triqs.gf import *\n",
    "from triqs.gf.descriptors import MatsubaraToLegendre"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Loading DB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2000\n"
     ]
    }
   ],
   "source": [
    "def load_db(filename):\n",
    "    n_entries = 0\n",
    "    params = 0\n",
    "    smpls_list = 0\n",
    "    with HDFArchive(filename, 'r') as ar:\n",
    "        n_entries = ar['n_entries']\n",
    "        params = ar[\"params\"] \n",
    "        smpls_list = ar[\"samples\"] \n",
    "    # print(params)\n",
    "    data2train = [ [[sample['time2solve'], sample['averageorder']],\n",
    "                    [params[smpl_id]['U'], params[smpl_id]['beta'], \n",
    "                      np.real(sample['G0_l'].data).flatten()[:30]], \n",
    "                    [ np.real(sample['G_l'].data).flatten()]]  \n",
    "                  for smpl_id, sample in enumerate(smpls_list)] \n",
    "    return data2train\n",
    "\n",
    "# db_names = [\n",
    "#     'database_v4.h5',\n",
    "#     'database_v5.h5',\n",
    "#     'database_v6.h5'\n",
    "#             ]\n",
    "db_names = ['./data/databaseV8.h5',\n",
    "            './data/databaseV9.h5',\n",
    "            './data/databaseV10.h5',\n",
    "            './data/databaseV11.h5']\n",
    "data_list = [load_db(db_name) for db_name in db_names]\n",
    "data = []\n",
    "for dataq in data_list:\n",
    "    data.extend(dataq)\n",
    "print(len(data))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "U       = np.array([sample[1][0] for sample in data])\n",
    "beta    = np.array([sample[1][1] for sample in data])\n",
    "times   = np.array([sample[0][0] for sample in data])\n",
    "order   = np.array([sample[0][1] for sample in data])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x250 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax, bx, cx) = plt.subplots(  1, 3, figsize=( 8, 2.5), dpi=100)\n",
    "\n",
    "ax.scatter(beta, times, alpha=0.2)\n",
    "ax.set_xlabel(r'U $\\beta$')\n",
    "ax.set_ylabel(r'$t$')\n",
    "\n",
    "# bx.scatter(U, time)\n",
    "bx.scatter(U, order, alpha=0.1)\n",
    "bx.set_xlabel(r'$U beta$')\n",
    "bx.set_ylabel('order')\n",
    "\n",
    "cx.scatter(order, times, alpha=0.1)\n",
    "cx.set_xlabel(r'$order$')\n",
    "cx.set_ylabel(r'$t$')\n",
    "\n",
    "\n",
    "fig.tight_layout()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Designing NN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "in total 2000 rows \n",
      "\n",
      "\n",
      "num features 32\n",
      "num targets 30\n",
      "num params 2\n",
      "\n",
      "\n",
      "num train 1400\n",
      "num val 402\n",
      "num test 198\n"
     ]
    }
   ],
   "source": [
    "def make_training_Data(db):\n",
    "    X = np.array([np.hstack(sample[1]) for sample in db])\n",
    "    Y = np.array([sample[2][0] for sample in db])\n",
    "    params = np.array([sample[0] for sample in db])\n",
    "    return X, Y, params\n",
    "\n",
    "features, targets, params = make_training_Data(data)\n",
    "print(f'in total {len(features)} rows \\n\\n')\n",
    "print(f'num features {len(features[0])}')\n",
    "print(f'num targets {len(targets[0])}')\n",
    "print(f'num params {len(params[0])}\\n\\n')\n",
    "\n",
    "X_train, X_test, y_train, y_test, params_train, params_test = \\\n",
    "    train_test_split(features, targets, params, test_size=0.3, random_state=42)\n",
    "\n",
    "X_val, X_test, y_val, y_test, params_val, params_test = \\\n",
    "    train_test_split(X_test, y_test, params_test, test_size=0.33, random_state=42)\n",
    "\n",
    "\n",
    "print(f'num train {len(X_train)}')\n",
    "print(f'num val {len(X_val)}')\n",
    "print(f'num test {len(X_test)}')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "class TinyNeuralNetwork(nn.Module):\n",
    "    def __init__(self, input_size=32, output_size=30):\n",
    "        super(self.__class__, self).__init__()\n",
    "        self.model = nn.Sequential(\n",
    "            # Your network structure comes here\n",
    "            nn.Linear(input_size, 64),  # First hidden layer\n",
    "            nn.ReLU(),                    # Activation function\n",
    "            nn.Linear(64, 128),           # Second hidden layer\n",
    "            nn.ReLU(),                    # Activation function\n",
    "            nn.Linear(128, 64),           # Third hidden layer\n",
    "            nn.ReLU(),                    # Activation function\n",
    "            nn.Linear(64, output_size)    # Output layer\n",
    "        )    \n",
    "        #     nn.Linear(784, 32),\n",
    "        #     nn.BatchNorm1d(32),\n",
    "        #     nn.ELU(),\n",
    "        #     nn.Linear(32,10),\n",
    "        #     nn.Linear(input_shape, num_classes)\n",
    "        # )\n",
    "        \n",
    "    def forward(self, inp):       \n",
    "        out = self.model(inp)\n",
    "        # print(out)\n",
    "        return out\n",
    "    \n",
    "# torchsummary.summary(TinyNeuralNetwork().to(device), (500, 42))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# batch generator\n",
    "def get_batches(dataset, batch_size):\n",
    "    X, Y, params = dataset\n",
    "    n_samples = X.shape[0]\n",
    "        \n",
    "    # Shuffle at the start of epoch\n",
    "    indices = np.arange(n_samples)\n",
    "    np.random.shuffle(indices)\n",
    "    \n",
    "    for start in range(0, n_samples, batch_size):\n",
    "        end = min(start + batch_size, n_samples)\n",
    "        \n",
    "        batch_idx = indices[start:end]\n",
    "    \n",
    "        yield torch.Tensor(X[batch_idx]), torch.Tensor(Y[batch_idx]), params[batch_idx]\n",
    "\n",
    "\n",
    "# batch_size = 4\n",
    "# train_loader = get_batches((X_train, y_train), batch_size)\n",
    "# test_loader = get_batches((X_test, y_test), batch_size)\n",
    "\n",
    "# batch_size = 128\n",
    "# for x_batch, y_batch in get_batches((X_train, y_train), batch_size):\n",
    "#     print(x_batch.shape)\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_model(model, batch_size, loss_func, opt, acc, n_epochs: int):\n",
    "    train_loss = []\n",
    "    val_loss = []\n",
    "    val_accuracy = []\n",
    "    \n",
    "    for epoch in range(n_epochs):\n",
    "        ep_train_loss = []\n",
    "        ep_val_loss = []\n",
    "        ep_val_accuracy = []\n",
    "        start_time = time.time()\n",
    "\n",
    "        model.train(True) # enable dropout / batch_norm training behavior\n",
    "        for X_batch, y_batch, _ in get_batches((X_train, y_train, params_train), batch_size):\n",
    "            # print(X_batch.shape)\n",
    "            \n",
    "            X_batch = X_batch.to(device, torch.float32)\n",
    "            y_batch = y_batch.to(device, torch.float32)\n",
    "            # train on batch: compute loss, calc grads, perform optimizer step and zero the grads\n",
    "            \n",
    "            # print(X_batch)\n",
    "            y_pred = model(X_batch)\n",
    "            # print(y_pred)\n",
    "            loss = loss_func(y_pred, y_batch)\n",
    "            loss.backward()\n",
    "            opt.step()\n",
    "            opt.zero_grad()\n",
    "\n",
    "            ep_train_loss.append(loss.item())\n",
    "\n",
    "        \n",
    "        model.train(False) # disable dropout / use averages for batch_norm\n",
    "        with torch.no_grad():\n",
    "            for X_batch, y_batch, _ in get_batches((X_val, y_val, params_val), batch_size):\n",
    "                # move data to target device\n",
    "                \n",
    "                X_batch = X_batch.to(device, torch.float32)\n",
    "                y_batch = y_batch.to(device, torch.float32)\n",
    "                # compute predictions\n",
    "                \n",
    "                y_pred = model(X_batch)\n",
    "                ep_val_loss.append( loss_func(y_pred, y_batch) )\n",
    "                ep_val_accuracy.append( acc(y_pred, y_batch))\n",
    "\n",
    "        # print the results for this epoch:\n",
    "        print(f'Epoch {epoch + 1} of {n_epochs} took {time.time() - start_time:.3f}s')\n",
    "\n",
    "        train_loss.append(np.mean(ep_train_loss))\n",
    "        val_loss.append(np.mean(ep_val_loss))\n",
    "        val_accuracy.append(np.mean(ep_val_accuracy))\n",
    "        \n",
    "        print(f\"\\t  training loss: {train_loss[-1]:.6f}\")\n",
    "        print(f\"\\tvalidation loss: {val_loss[-1]:.6f}\")\n",
    "        print(f\"\\tvalidation accuracy: {val_accuracy[-1]:.3f}\")\n",
    "\n",
    "    return train_loss, val_loss, val_accuracy\n",
    "\n",
    "\n",
    "def save_checkpoint(checkpoint_path, model, optimizer):\n",
    "    # state_dict: a Python dictionary object that:\n",
    "    # - for a model, maps each layer to its parameter tensor;\n",
    "    # - for an optimizer, contains info about the optimizer’s states and hyperparameters used.\n",
    "    state = {\n",
    "        'state_dict': model.state_dict(),\n",
    "        'optimizer' : optimizer.state_dict()}\n",
    "    torch.save(state, checkpoint_path)\n",
    "    print('model saved to %s' % checkpoint_path)\n",
    "    \n",
    "def load_checkpoint(checkpoint_path, model, optimizer):\n",
    "    state = torch.load(checkpoint_path)\n",
    "    model.load_state_dict(state['state_dict'])\n",
    "    optimizer.load_state_dict(state['optimizer'])\n",
    "    print('model loaded from %s' % checkpoint_path)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training NN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = TinyNeuralNetwork().to(device, torch.float32)\n",
    "opt =  torch.optim.Adam(model.parameters(), lr=1e-3) \n",
    "loss_func = nn.MSELoss()\n",
    "acc = torchmetrics.MeanSquaredError()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1 of 200 took 0.031s\n",
      "\t  training loss: 0.000567\n",
      "\tvalidation loss: 0.000603\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 2 of 200 took 0.026s\n",
      "\t  training loss: 0.000567\n",
      "\tvalidation loss: 0.000546\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 3 of 200 took 0.025s\n",
      "\t  training loss: 0.000569\n",
      "\tvalidation loss: 0.000568\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 4 of 200 took 0.022s\n",
      "\t  training loss: 0.000573\n",
      "\tvalidation loss: 0.000608\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 5 of 200 took 0.022s\n",
      "\t  training loss: 0.000574\n",
      "\tvalidation loss: 0.000590\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 6 of 200 took 0.021s\n",
      "\t  training loss: 0.000578\n",
      "\tvalidation loss: 0.000584\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 7 of 200 took 0.020s\n",
      "\t  training loss: 0.000571\n",
      "\tvalidation loss: 0.000696\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 8 of 200 took 0.021s\n",
      "\t  training loss: 0.000567\n",
      "\tvalidation loss: 0.000538\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 9 of 200 took 0.020s\n",
      "\t  training loss: 0.000574\n",
      "\tvalidation loss: 0.000626\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 10 of 200 took 0.020s\n",
      "\t  training loss: 0.000560\n",
      "\tvalidation loss: 0.000618\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 11 of 200 took 0.021s\n",
      "\t  training loss: 0.000572\n",
      "\tvalidation loss: 0.000599\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 12 of 200 took 0.021s\n",
      "\t  training loss: 0.000566\n",
      "\tvalidation loss: 0.000630\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 13 of 200 took 0.022s\n",
      "\t  training loss: 0.000562\n",
      "\tvalidation loss: 0.000559\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 14 of 200 took 0.021s\n",
      "\t  training loss: 0.000578\n",
      "\tvalidation loss: 0.000603\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 15 of 200 took 0.018s\n",
      "\t  training loss: 0.000569\n",
      "\tvalidation loss: 0.000546\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 16 of 200 took 0.018s\n",
      "\t  training loss: 0.000560\n",
      "\tvalidation loss: 0.000642\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 17 of 200 took 0.019s\n",
      "\t  training loss: 0.000559\n",
      "\tvalidation loss: 0.000588\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 18 of 200 took 0.021s\n",
      "\t  training loss: 0.000555\n",
      "\tvalidation loss: 0.000565\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 19 of 200 took 0.018s\n",
      "\t  training loss: 0.000564\n",
      "\tvalidation loss: 0.000595\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 20 of 200 took 0.018s\n",
      "\t  training loss: 0.000570\n",
      "\tvalidation loss: 0.000561\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 21 of 200 took 0.018s\n",
      "\t  training loss: 0.000575\n",
      "\tvalidation loss: 0.000558\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 22 of 200 took 0.020s\n",
      "\t  training loss: 0.000570\n",
      "\tvalidation loss: 0.000639\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 23 of 200 took 0.024s\n",
      "\t  training loss: 0.000570\n",
      "\tvalidation loss: 0.000551\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 24 of 200 took 0.029s\n",
      "\t  training loss: 0.000558\n",
      "\tvalidation loss: 0.000529\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 25 of 200 took 0.033s\n",
      "\t  training loss: 0.000554\n",
      "\tvalidation loss: 0.000632\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 26 of 200 took 0.027s\n",
      "\t  training loss: 0.000551\n",
      "\tvalidation loss: 0.000551\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 27 of 200 took 0.029s\n",
      "\t  training loss: 0.000551\n",
      "\tvalidation loss: 0.000550\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 28 of 200 took 0.030s\n",
      "\t  training loss: 0.000555\n",
      "\tvalidation loss: 0.000562\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 29 of 200 took 0.025s\n",
      "\t  training loss: 0.000554\n",
      "\tvalidation loss: 0.000601\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 30 of 200 took 0.023s\n",
      "\t  training loss: 0.000559\n",
      "\tvalidation loss: 0.000553\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 31 of 200 took 0.020s\n",
      "\t  training loss: 0.000552\n",
      "\tvalidation loss: 0.000600\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 32 of 200 took 0.019s\n",
      "\t  training loss: 0.000560\n",
      "\tvalidation loss: 0.000656\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 33 of 200 took 0.021s\n",
      "\t  training loss: 0.000561\n",
      "\tvalidation loss: 0.000593\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 34 of 200 took 0.019s\n",
      "\t  training loss: 0.000565\n",
      "\tvalidation loss: 0.000575\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 35 of 200 took 0.022s\n",
      "\t  training loss: 0.000561\n",
      "\tvalidation loss: 0.000575\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 36 of 200 took 0.019s\n",
      "\t  training loss: 0.000556\n",
      "\tvalidation loss: 0.000584\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 37 of 200 took 0.018s\n",
      "\t  training loss: 0.000557\n",
      "\tvalidation loss: 0.000532\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 38 of 200 took 0.018s\n",
      "\t  training loss: 0.000561\n",
      "\tvalidation loss: 0.000585\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 39 of 200 took 0.019s\n",
      "\t  training loss: 0.000556\n",
      "\tvalidation loss: 0.000570\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 40 of 200 took 0.019s\n",
      "\t  training loss: 0.000554\n",
      "\tvalidation loss: 0.000518\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 41 of 200 took 0.021s\n",
      "\t  training loss: 0.000544\n",
      "\tvalidation loss: 0.000524\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 42 of 200 took 0.019s\n",
      "\t  training loss: 0.000549\n",
      "\tvalidation loss: 0.000535\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 43 of 200 took 0.019s\n",
      "\t  training loss: 0.000546\n",
      "\tvalidation loss: 0.000528\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 44 of 200 took 0.019s\n",
      "\t  training loss: 0.000550\n",
      "\tvalidation loss: 0.000542\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 45 of 200 took 0.019s\n",
      "\t  training loss: 0.000552\n",
      "\tvalidation loss: 0.000591\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 46 of 200 took 0.019s\n",
      "\t  training loss: 0.000553\n",
      "\tvalidation loss: 0.000584\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 47 of 200 took 0.021s\n",
      "\t  training loss: 0.000543\n",
      "\tvalidation loss: 0.000569\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 48 of 200 took 0.019s\n",
      "\t  training loss: 0.000563\n",
      "\tvalidation loss: 0.000604\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 49 of 200 took 0.019s\n",
      "\t  training loss: 0.000548\n",
      "\tvalidation loss: 0.000542\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 50 of 200 took 0.019s\n",
      "\t  training loss: 0.000547\n",
      "\tvalidation loss: 0.000588\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 51 of 200 took 0.019s\n",
      "\t  training loss: 0.000562\n",
      "\tvalidation loss: 0.000560\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 52 of 200 took 0.022s\n",
      "\t  training loss: 0.000538\n",
      "\tvalidation loss: 0.000484\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 53 of 200 took 0.020s\n",
      "\t  training loss: 0.000541\n",
      "\tvalidation loss: 0.000505\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 54 of 200 took 0.018s\n",
      "\t  training loss: 0.000542\n",
      "\tvalidation loss: 0.000528\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 55 of 200 took 0.018s\n",
      "\t  training loss: 0.000536\n",
      "\tvalidation loss: 0.000618\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 56 of 200 took 0.018s\n",
      "\t  training loss: 0.000538\n",
      "\tvalidation loss: 0.000583\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 57 of 200 took 0.019s\n",
      "\t  training loss: 0.000538\n",
      "\tvalidation loss: 0.000538\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 58 of 200 took 0.018s\n",
      "\t  training loss: 0.000541\n",
      "\tvalidation loss: 0.000519\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 59 of 200 took 0.019s\n",
      "\t  training loss: 0.000535\n",
      "\tvalidation loss: 0.000580\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 60 of 200 took 0.019s\n",
      "\t  training loss: 0.000540\n",
      "\tvalidation loss: 0.000536\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 61 of 200 took 0.019s\n",
      "\t  training loss: 0.000538\n",
      "\tvalidation loss: 0.000555\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 62 of 200 took 0.019s\n",
      "\t  training loss: 0.000550\n",
      "\tvalidation loss: 0.000575\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 63 of 200 took 0.019s\n",
      "\t  training loss: 0.000540\n",
      "\tvalidation loss: 0.000505\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 64 of 200 took 0.020s\n",
      "\t  training loss: 0.000537\n",
      "\tvalidation loss: 0.000552\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 65 of 200 took 0.020s\n",
      "\t  training loss: 0.000538\n",
      "\tvalidation loss: 0.000560\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 66 of 200 took 0.023s\n",
      "\t  training loss: 0.000536\n",
      "\tvalidation loss: 0.000542\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 67 of 200 took 0.021s\n",
      "\t  training loss: 0.000534\n",
      "\tvalidation loss: 0.000497\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 68 of 200 took 0.020s\n",
      "\t  training loss: 0.000530\n",
      "\tvalidation loss: 0.000546\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 69 of 200 took 0.020s\n",
      "\t  training loss: 0.000536\n",
      "\tvalidation loss: 0.000576\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 70 of 200 took 0.020s\n",
      "\t  training loss: 0.000541\n",
      "\tvalidation loss: 0.000505\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 71 of 200 took 0.019s\n",
      "\t  training loss: 0.000545\n",
      "\tvalidation loss: 0.000542\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 72 of 200 took 0.019s\n",
      "\t  training loss: 0.000549\n",
      "\tvalidation loss: 0.000551\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 73 of 200 took 0.019s\n",
      "\t  training loss: 0.000536\n",
      "\tvalidation loss: 0.000511\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 74 of 200 took 0.022s\n",
      "\t  training loss: 0.000536\n",
      "\tvalidation loss: 0.000511\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 75 of 200 took 0.019s\n",
      "\t  training loss: 0.000542\n",
      "\tvalidation loss: 0.000500\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 76 of 200 took 0.019s\n",
      "\t  training loss: 0.000544\n",
      "\tvalidation loss: 0.000599\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 77 of 200 took 0.019s\n",
      "\t  training loss: 0.000529\n",
      "\tvalidation loss: 0.000537\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 78 of 200 took 0.019s\n",
      "\t  training loss: 0.000527\n",
      "\tvalidation loss: 0.000540\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 79 of 200 took 0.018s\n",
      "\t  training loss: 0.000529\n",
      "\tvalidation loss: 0.000535\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 80 of 200 took 0.019s\n",
      "\t  training loss: 0.000533\n",
      "\tvalidation loss: 0.000527\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 81 of 200 took 0.020s\n",
      "\t  training loss: 0.000542\n",
      "\tvalidation loss: 0.000560\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 82 of 200 took 0.019s\n",
      "\t  training loss: 0.000531\n",
      "\tvalidation loss: 0.000571\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 83 of 200 took 0.019s\n",
      "\t  training loss: 0.000536\n",
      "\tvalidation loss: 0.000601\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 84 of 200 took 0.019s\n",
      "\t  training loss: 0.000532\n",
      "\tvalidation loss: 0.000587\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 85 of 200 took 0.019s\n",
      "\t  training loss: 0.000538\n",
      "\tvalidation loss: 0.000558\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 86 of 200 took 0.019s\n",
      "\t  training loss: 0.000526\n",
      "\tvalidation loss: 0.000549\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 87 of 200 took 0.020s\n",
      "\t  training loss: 0.000533\n",
      "\tvalidation loss: 0.000564\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 88 of 200 took 0.019s\n",
      "\t  training loss: 0.000540\n",
      "\tvalidation loss: 0.000495\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 89 of 200 took 0.019s\n",
      "\t  training loss: 0.000542\n",
      "\tvalidation loss: 0.000539\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 90 of 200 took 0.018s\n",
      "\t  training loss: 0.000533\n",
      "\tvalidation loss: 0.000505\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 91 of 200 took 0.019s\n",
      "\t  training loss: 0.000542\n",
      "\tvalidation loss: 0.000596\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 92 of 200 took 0.018s\n",
      "\t  training loss: 0.000543\n",
      "\tvalidation loss: 0.000626\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 93 of 200 took 0.019s\n",
      "\t  training loss: 0.000539\n",
      "\tvalidation loss: 0.000604\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 94 of 200 took 0.019s\n",
      "\t  training loss: 0.000540\n",
      "\tvalidation loss: 0.000522\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 95 of 200 took 0.020s\n",
      "\t  training loss: 0.000531\n",
      "\tvalidation loss: 0.000501\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 96 of 200 took 0.020s\n",
      "\t  training loss: 0.000533\n",
      "\tvalidation loss: 0.000547\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 97 of 200 took 0.018s\n",
      "\t  training loss: 0.000529\n",
      "\tvalidation loss: 0.000540\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 98 of 200 took 0.019s\n",
      "\t  training loss: 0.000537\n",
      "\tvalidation loss: 0.000515\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 99 of 200 took 0.019s\n",
      "\t  training loss: 0.000526\n",
      "\tvalidation loss: 0.000504\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 100 of 200 took 0.019s\n",
      "\t  training loss: 0.000518\n",
      "\tvalidation loss: 0.000601\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 101 of 200 took 0.020s\n",
      "\t  training loss: 0.000524\n",
      "\tvalidation loss: 0.000536\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 102 of 200 took 0.019s\n",
      "\t  training loss: 0.000534\n",
      "\tvalidation loss: 0.000531\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 103 of 200 took 0.018s\n",
      "\t  training loss: 0.000530\n",
      "\tvalidation loss: 0.000561\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 104 of 200 took 0.018s\n",
      "\t  training loss: 0.000529\n",
      "\tvalidation loss: 0.000618\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 105 of 200 took 0.018s\n",
      "\t  training loss: 0.000524\n",
      "\tvalidation loss: 0.000503\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 106 of 200 took 0.019s\n",
      "\t  training loss: 0.000529\n",
      "\tvalidation loss: 0.000625\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 107 of 200 took 0.021s\n",
      "\t  training loss: 0.000526\n",
      "\tvalidation loss: 0.000617\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 108 of 200 took 0.020s\n",
      "\t  training loss: 0.000527\n",
      "\tvalidation loss: 0.000582\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 109 of 200 took 0.019s\n",
      "\t  training loss: 0.000531\n",
      "\tvalidation loss: 0.000590\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 110 of 200 took 0.019s\n",
      "\t  training loss: 0.000531\n",
      "\tvalidation loss: 0.000531\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 111 of 200 took 0.018s\n",
      "\t  training loss: 0.000532\n",
      "\tvalidation loss: 0.000561\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 112 of 200 took 0.018s\n",
      "\t  training loss: 0.000527\n",
      "\tvalidation loss: 0.000573\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 113 of 200 took 0.018s\n",
      "\t  training loss: 0.000526\n",
      "\tvalidation loss: 0.000539\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 114 of 200 took 0.019s\n",
      "\t  training loss: 0.000534\n",
      "\tvalidation loss: 0.000532\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 115 of 200 took 0.020s\n",
      "\t  training loss: 0.000531\n",
      "\tvalidation loss: 0.000606\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 116 of 200 took 0.020s\n",
      "\t  training loss: 0.000532\n",
      "\tvalidation loss: 0.000572\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 117 of 200 took 0.021s\n",
      "\t  training loss: 0.000531\n",
      "\tvalidation loss: 0.000492\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 118 of 200 took 0.028s\n",
      "\t  training loss: 0.000523\n",
      "\tvalidation loss: 0.000577\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 119 of 200 took 0.020s\n",
      "\t  training loss: 0.000527\n",
      "\tvalidation loss: 0.000531\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 120 of 200 took 0.023s\n",
      "\t  training loss: 0.000535\n",
      "\tvalidation loss: 0.000569\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 121 of 200 took 0.027s\n",
      "\t  training loss: 0.000530\n",
      "\tvalidation loss: 0.000527\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 122 of 200 took 0.021s\n",
      "\t  training loss: 0.000526\n",
      "\tvalidation loss: 0.000540\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 123 of 200 took 0.020s\n",
      "\t  training loss: 0.000526\n",
      "\tvalidation loss: 0.000515\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 124 of 200 took 0.022s\n",
      "\t  training loss: 0.000529\n",
      "\tvalidation loss: 0.000504\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 125 of 200 took 0.020s\n",
      "\t  training loss: 0.000525\n",
      "\tvalidation loss: 0.000521\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 126 of 200 took 0.019s\n",
      "\t  training loss: 0.000530\n",
      "\tvalidation loss: 0.000517\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 127 of 200 took 0.021s\n",
      "\t  training loss: 0.000521\n",
      "\tvalidation loss: 0.000549\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 128 of 200 took 0.019s\n",
      "\t  training loss: 0.000521\n",
      "\tvalidation loss: 0.000536\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 129 of 200 took 0.019s\n",
      "\t  training loss: 0.000519\n",
      "\tvalidation loss: 0.000576\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 130 of 200 took 0.019s\n",
      "\t  training loss: 0.000527\n",
      "\tvalidation loss: 0.000558\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 131 of 200 took 0.019s\n",
      "\t  training loss: 0.000518\n",
      "\tvalidation loss: 0.000554\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 132 of 200 took 0.018s\n",
      "\t  training loss: 0.000530\n",
      "\tvalidation loss: 0.000515\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 133 of 200 took 0.018s\n",
      "\t  training loss: 0.000526\n",
      "\tvalidation loss: 0.000502\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 134 of 200 took 0.018s\n",
      "\t  training loss: 0.000522\n",
      "\tvalidation loss: 0.000495\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 135 of 200 took 0.020s\n",
      "\t  training loss: 0.000523\n",
      "\tvalidation loss: 0.000479\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 136 of 200 took 0.020s\n",
      "\t  training loss: 0.000529\n",
      "\tvalidation loss: 0.000538\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 137 of 200 took 0.019s\n",
      "\t  training loss: 0.000517\n",
      "\tvalidation loss: 0.000523\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 138 of 200 took 0.020s\n",
      "\t  training loss: 0.000516\n",
      "\tvalidation loss: 0.000560\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 139 of 200 took 0.017s\n",
      "\t  training loss: 0.000540\n",
      "\tvalidation loss: 0.000554\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 140 of 200 took 0.018s\n",
      "\t  training loss: 0.000527\n",
      "\tvalidation loss: 0.000502\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 141 of 200 took 0.017s\n",
      "\t  training loss: 0.000523\n",
      "\tvalidation loss: 0.000529\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 142 of 200 took 0.020s\n",
      "\t  training loss: 0.000520\n",
      "\tvalidation loss: 0.000485\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 143 of 200 took 0.017s\n",
      "\t  training loss: 0.000515\n",
      "\tvalidation loss: 0.000510\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 144 of 200 took 0.018s\n",
      "\t  training loss: 0.000513\n",
      "\tvalidation loss: 0.000527\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 145 of 200 took 0.018s\n",
      "\t  training loss: 0.000516\n",
      "\tvalidation loss: 0.000512\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 146 of 200 took 0.017s\n",
      "\t  training loss: 0.000513\n",
      "\tvalidation loss: 0.000562\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 147 of 200 took 0.020s\n",
      "\t  training loss: 0.000522\n",
      "\tvalidation loss: 0.000532\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 148 of 200 took 0.018s\n",
      "\t  training loss: 0.000533\n",
      "\tvalidation loss: 0.000553\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 149 of 200 took 0.020s\n",
      "\t  training loss: 0.000527\n",
      "\tvalidation loss: 0.000546\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 150 of 200 took 0.019s\n",
      "\t  training loss: 0.000517\n",
      "\tvalidation loss: 0.000545\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 151 of 200 took 0.020s\n",
      "\t  training loss: 0.000519\n",
      "\tvalidation loss: 0.000530\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 152 of 200 took 0.019s\n",
      "\t  training loss: 0.000521\n",
      "\tvalidation loss: 0.000559\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 153 of 200 took 0.019s\n",
      "\t  training loss: 0.000528\n",
      "\tvalidation loss: 0.000496\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 154 of 200 took 0.018s\n",
      "\t  training loss: 0.000513\n",
      "\tvalidation loss: 0.000524\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 155 of 200 took 0.018s\n",
      "\t  training loss: 0.000513\n",
      "\tvalidation loss: 0.000497\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 156 of 200 took 0.019s\n",
      "\t  training loss: 0.000522\n",
      "\tvalidation loss: 0.000573\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 157 of 200 took 0.018s\n",
      "\t  training loss: 0.000527\n",
      "\tvalidation loss: 0.000499\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 158 of 200 took 0.019s\n",
      "\t  training loss: 0.000535\n",
      "\tvalidation loss: 0.000579\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 159 of 200 took 0.020s\n",
      "\t  training loss: 0.000534\n",
      "\tvalidation loss: 0.000568\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 160 of 200 took 0.020s\n",
      "\t  training loss: 0.000536\n",
      "\tvalidation loss: 0.000536\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 161 of 200 took 0.018s\n",
      "\t  training loss: 0.000521\n",
      "\tvalidation loss: 0.000522\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 162 of 200 took 0.018s\n",
      "\t  training loss: 0.000512\n",
      "\tvalidation loss: 0.000557\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 163 of 200 took 0.019s\n",
      "\t  training loss: 0.000519\n",
      "\tvalidation loss: 0.000521\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 164 of 200 took 0.018s\n",
      "\t  training loss: 0.000518\n",
      "\tvalidation loss: 0.000528\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 165 of 200 took 0.018s\n",
      "\t  training loss: 0.000516\n",
      "\tvalidation loss: 0.000503\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 166 of 200 took 0.019s\n",
      "\t  training loss: 0.000512\n",
      "\tvalidation loss: 0.000552\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 167 of 200 took 0.020s\n",
      "\t  training loss: 0.000509\n",
      "\tvalidation loss: 0.000520\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 168 of 200 took 0.019s\n",
      "\t  training loss: 0.000518\n",
      "\tvalidation loss: 0.000512\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 169 of 200 took 0.021s\n",
      "\t  training loss: 0.000526\n",
      "\tvalidation loss: 0.000467\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 170 of 200 took 0.020s\n",
      "\t  training loss: 0.000508\n",
      "\tvalidation loss: 0.000542\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 171 of 200 took 0.022s\n",
      "\t  training loss: 0.000513\n",
      "\tvalidation loss: 0.000509\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 172 of 200 took 0.019s\n",
      "\t  training loss: 0.000508\n",
      "\tvalidation loss: 0.000539\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 173 of 200 took 0.019s\n",
      "\t  training loss: 0.000504\n",
      "\tvalidation loss: 0.000526\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 174 of 200 took 0.018s\n",
      "\t  training loss: 0.000513\n",
      "\tvalidation loss: 0.000516\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 175 of 200 took 0.018s\n",
      "\t  training loss: 0.000513\n",
      "\tvalidation loss: 0.000543\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 176 of 200 took 0.018s\n",
      "\t  training loss: 0.000513\n",
      "\tvalidation loss: 0.000544\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 177 of 200 took 0.019s\n",
      "\t  training loss: 0.000511\n",
      "\tvalidation loss: 0.000520\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 178 of 200 took 0.021s\n",
      "\t  training loss: 0.000519\n",
      "\tvalidation loss: 0.000506\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 179 of 200 took 0.020s\n",
      "\t  training loss: 0.000515\n",
      "\tvalidation loss: 0.000536\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 180 of 200 took 0.018s\n",
      "\t  training loss: 0.000513\n",
      "\tvalidation loss: 0.000495\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 181 of 200 took 0.019s\n",
      "\t  training loss: 0.000507\n",
      "\tvalidation loss: 0.000517\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 182 of 200 took 0.020s\n",
      "\t  training loss: 0.000514\n",
      "\tvalidation loss: 0.000539\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 183 of 200 took 0.019s\n",
      "\t  training loss: 0.000531\n",
      "\tvalidation loss: 0.000491\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 184 of 200 took 0.021s\n",
      "\t  training loss: 0.000521\n",
      "\tvalidation loss: 0.000510\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 185 of 200 took 0.018s\n",
      "\t  training loss: 0.000512\n",
      "\tvalidation loss: 0.000567\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 186 of 200 took 0.019s\n",
      "\t  training loss: 0.000512\n",
      "\tvalidation loss: 0.000555\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 187 of 200 took 0.018s\n",
      "\t  training loss: 0.000509\n",
      "\tvalidation loss: 0.000556\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 188 of 200 took 0.019s\n",
      "\t  training loss: 0.000509\n",
      "\tvalidation loss: 0.000527\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 189 of 200 took 0.019s\n",
      "\t  training loss: 0.000511\n",
      "\tvalidation loss: 0.000476\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 190 of 200 took 0.019s\n",
      "\t  training loss: 0.000518\n",
      "\tvalidation loss: 0.000500\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 191 of 200 took 0.020s\n",
      "\t  training loss: 0.000516\n",
      "\tvalidation loss: 0.000543\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 192 of 200 took 0.019s\n",
      "\t  training loss: 0.000501\n",
      "\tvalidation loss: 0.000522\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 193 of 200 took 0.019s\n",
      "\t  training loss: 0.000508\n",
      "\tvalidation loss: 0.000562\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 194 of 200 took 0.019s\n",
      "\t  training loss: 0.000520\n",
      "\tvalidation loss: 0.000528\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 195 of 200 took 0.018s\n",
      "\t  training loss: 0.000518\n",
      "\tvalidation loss: 0.000587\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 196 of 200 took 0.018s\n",
      "\t  training loss: 0.000510\n",
      "\tvalidation loss: 0.000509\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 197 of 200 took 0.020s\n",
      "\t  training loss: 0.000505\n",
      "\tvalidation loss: 0.000489\n",
      "\tvalidation accuracy: 0.000\n",
      "Epoch 198 of 200 took 0.018s\n",
      "\t  training loss: 0.000503\n",
      "\tvalidation loss: 0.000542\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 199 of 200 took 0.019s\n",
      "\t  training loss: 0.000508\n",
      "\tvalidation loss: 0.000503\n",
      "\tvalidation accuracy: 0.001\n",
      "Epoch 200 of 200 took 0.018s\n",
      "\t  training loss: 0.000502\n",
      "\tvalidation loss: 0.000482\n",
      "\tvalidation accuracy: 0.000\n"
     ]
    }
   ],
   "source": [
    "n_epochs = 200\n",
    "batch_size = 128\n",
    "train_loss, test_loss, val_accuracy = train_model(model, batch_size, loss_func, opt, acc, n_epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "def plot_train_process(train_loss, test_loss, val_accuracy):\n",
    "    fig, ax = plt.subplots( figsize=(5, 4))\n",
    "\n",
    "    # ax.set_title('Loss')\n",
    "    ax.plot(train_loss, label='Train')\n",
    "    ax.plot(test_loss, label='Val')\n",
    "    ax.set_yscale('log')\n",
    "\n",
    "    ax.set_xlim((0, 100))\n",
    "    ax.set_xlabel('Epoch')\n",
    "    ax.set_ylabel('Loss')\n",
    "    fig.tight_layout()\n",
    "\n",
    "    ax.legend()\n",
    "\n",
    "plot_train_process(train_loss, test_loss, val_accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "model saved to ./models/model_v3\n"
     ]
    }
   ],
   "source": [
    "save_checkpoint(checkpoint_path='./models/model_v3', model=model, optimizer=opt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "model loaded from ./models/model_v3\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_18372/3687492894.py:69: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
      "  state = torch.load(checkpoint_path)\n"
     ]
    }
   ],
   "source": [
    "load_checkpoint(checkpoint_path='./models/model_v3', model=model, optimizer=opt)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Error analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "128 128\n",
      "70 70\n"
     ]
    }
   ],
   "source": [
    "batch_size = 128\n",
    "\n",
    "ep_test_accuracy = []\n",
    "model.train(False) # disable dropout / use averages for batch_norm\n",
    "with torch.no_grad():\n",
    "    for X_batch, y_batch, params_batch in get_batches((X_test, y_test, params_test), batch_size):\n",
    "        # move data to target device\n",
    "        \n",
    "        X_batch = X_batch.to(device, torch.float32)\n",
    "        y_batch = y_batch.to(device, torch.float32)\n",
    "        # compute predictions\n",
    "        \n",
    "        y_pred = model(X_batch)\n",
    "        print(len(params_batch), len(y_batch))\n",
    "        ep_test_accuracy.append( np.concatenate( (  \n",
    "            np.log(np.reshape([acc(y_pred[i], y_batch[i]).item() for i in range(len(y_batch)) ], (-1, 1) ) ), \n",
    "                                                  params_batch ), axis=1))\n",
    "\n",
    "test_mse = np.concatenate(ep_test_accuracy, axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "198"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# test_mse = pd.DataFrame({\n",
    "#     'mse': np.random.rand(100),\n",
    "#     'time': np.random.rand(100) * 100,\n",
    "#     'order': np.random.randint(1, 10, size=100)\n",
    "# })\n",
    "\n",
    "# Define the number of bins for the histogram\n",
    "num_bins = 10\n",
    "\n",
    "# Create the histogram for MSE\n",
    "counts, bin_edges = np.histogram(test_mse[:, 0], bins=num_bins)\n",
    "\n",
    "# Calculate the bin indices for each data point\n",
    "bin_indices = np.digitize(test_mse[:, 0], bin_edges)\n",
    "\n",
    "# Calculate the mean 'order' and 'time' for each bin\n",
    "mean_order = []\n",
    "mean_time = []\n",
    "\n",
    "for i in range(1, len(bin_edges)):\n",
    "    # Get the data points corresponding to the current bin\n",
    "    bin_data = test_mse[bin_indices == i]\n",
    "    if len(bin_data) > 0:\n",
    "        mean_order.append(bin_data[:, 2].mean())\n",
    "        mean_time.append(bin_data[:, 1].mean())\n",
    "    else:\n",
    "        mean_order.append(0)\n",
    "        mean_time.append(0)\n",
    "\n",
    "# # Plotting\n",
    "# fig, ax1 = plt.subplots( figsize=( 5,2.5), dpi=200)\n",
    "\n",
    "# # Histogram for MSE\n",
    "# ax1.hist(test_mse[:, 0], bins=num_bins, alpha=0.7, color='b')\n",
    "# ax1.set_xlabel('log MSE')\n",
    "# ax1.set_ylabel('Frequency (log MSE)', color='b')\n",
    "\n",
    "# # Secondary axis for mean order and time\n",
    "# ax2 = ax1.twinx()\n",
    "# ax2.plot(bin_edges[:-1] + 0.5*(bin_edges[1]-bin_edges[0]), mean_order, 'r-', marker='o', label='Mean Order', linewidth=2)\n",
    "# # ax2.plot(bin_edges[:-1], mean_time, 'g-', marker='x', label='Mean Time', linewidth=2)\n",
    "# ax2.set_ylabel(r'$\\langle n \\rangle$', color='red')\n",
    "\n",
    "# # Add legends and titles\n",
    "# # fig.suptitle('MSE Histogram with Mean Order and Time')\n",
    "# # ax2.legend(loc='upper right')\n",
    "# # plt.savefig('./errors.pdf', dpi=200, bbox_inches='tight')\n",
    "\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x400 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "num_bins = 10\n",
    "\n",
    "# Create the histogram for MSE\n",
    "counts, bin_edges = np.histogram(np.exp(test_mse[:, 0]), bins=num_bins)\n",
    "\n",
    "\n",
    "def plot_all():\n",
    "\n",
    "\n",
    "    fig = plt.figure(constrained_layout=True, figsize=( 4, 4))\n",
    "    # axs = fig.subplot_mosaic([['Left', 'TopRight'],['Left', 'BottomRight']],\n",
    "    #                         gridspec_kw={'width_ratios':[2, 1]})\n",
    "    \n",
    "    axs = fig.subplot_mosaic([['Top', 'Top'],['BottomLeft', 'BottomRight']],\n",
    "                            gridspec_kw={'width_ratios':[1, 1]})\n",
    "    # axs['Top'].set_title('Spectral function')\n",
    "\n",
    "\n",
    "    axs['Top'].hist(np.exp(test_mse[:, 0]), bins=num_bins, alpha=0.8, color='b')\n",
    "    axs['Top'].set_xscale('log')\n",
    "    \n",
    "    # axs['Top'].hist(test_mse[:, 0], bins=num_bins, alpha=0.8, color='b')\n",
    "    \n",
    "\n",
    "    axs['Top'].set_xlabel('MSE')\n",
    "    axs['Top'].set_ylabel('Counts', color='b')\n",
    "\n",
    "    # Secondary axis for mean order and time\n",
    "    ax2 = axs['Top'].twinx()\n",
    "    ax2.plot(bin_edges[:-1] , mean_order, \n",
    "                    'r-', marker='o', label='Mean Order', linewidth=2)\n",
    "    # ax2.plot(bin_edges[:-1], mean_time, 'g-', marker='x', label='Mean Time', linewidth=2)\n",
    "    ax2.set_ylabel(r'$\\langle n \\rangle$', color='red')\n",
    "\n",
    "    # axs['Top'].set_xlim((-1, 1))\n",
    "    # axs['Top'].set_ylim((0, max(np.max(A0)*1.5,  np.max(A0)*1.5) ))\n",
    "    # axs['Top'].set_xlabel(r'$E - E_f$(eV)')\n",
    "    # axs['Top'].set_ylabel('Density of States (DOS)')\n",
    "    axs['Top'].text(-0.15, 1.0, 'a)', transform=axs['Top'].transAxes,\n",
    "        fontsize=10, fontweight='normal', va='top', ha='right')\n",
    "    \n",
    "    # axs['Top'].legend(prop={'size': 10}, frameon=False)  \n",
    "\n",
    "    # plt.legend()\n",
    "\n",
    "    #################################################################################################    \n",
    "    # axs['BottomLeft'].set_title('Plot Top Right')\n",
    "    axs['BottomLeft'].hist(params_test[:, 1], density=False, alpha=0.8,  color='b')\n",
    "\n",
    "    axs['BottomLeft'].set_xlabel(r'$\\langle n \\rangle$')\n",
    "    axs['BottomLeft'].set_ylabel(r'Counts')\n",
    "    # axs['BottomLeft'].set_xlim((0, sample.beta))\n",
    "    \n",
    "    axs['BottomLeft'].text(-0.35, 1.0, 'b)', transform=axs['BottomLeft'].transAxes,\n",
    "        fontsize=10, fontweight='normal', va='top', ha='right')\n",
    "    \n",
    "    #################################################################################################    \n",
    "    # axs['BottomRight'].set_title('Legendre')\n",
    "    axs['BottomRight'].hist(X_test[:, 0], density=False,alpha=0.8,  color='b')\n",
    "    \n",
    "    axs['BottomRight'].set_xlabel(r'$U$, eV')\n",
    "    axs['BottomRight'].set_ylabel('Counts')\n",
    "    # axs['BottomRight'].set_xlim((0, 10))\n",
    "    \n",
    "    \n",
    "    axs['BottomRight'].text(-0.2, 1.0, 'c)', transform=axs['BottomRight'].transAxes,\n",
    "        fontsize=10, fontweight='normal', va='top', ha='right')\n",
    "    \n",
    "    \n",
    "    plt.savefig('./pics/dispersion.pdf', dpi=200, bbox_inches='tight')\n",
    "\n",
    "    plt.show()\n",
    "\n",
    "\n",
    "plot_all()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "triqs",
   "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": "3.12.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}