CGNN_oqmd_fe_benchmark_torch1102+cuda102.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyOIKvCEbHUUtSf6jB+VWi2z",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Tony-Y/oqmd-v1.2-dataset-for-cgnn/blob/main/CGNN_oqmd_fe_benchmark_torch1102%2Bcuda102.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rzcDEzcWsysZ",
        "outputId": "4f9bb0f1-e433-48b5-d9d8-c0a0cd0660f2"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "GPU 0: Tesla T4 (UUID: GPU-21eae000-c1c2-b6c2-8a98-7fe744c4ee57)\n"
          ]
        }
      ],
      "source": [
        "!nvidia-smi -L"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Install PyTorch 1.10.2 with CUDA 10.2"
      ],
      "metadata": {
        "id": "sWZ_P6ZxtL6C"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip uninstall -y torchvision torchaudio torchtext fastai"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ssELu3RTtI7C",
        "outputId": "3eef7f7a-681d-4069-a4eb-62cc19d5a5cf"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found existing installation: torchvision 0.13.1+cu113\n",
            "Uninstalling torchvision-0.13.1+cu113:\n",
            "  Successfully uninstalled torchvision-0.13.1+cu113\n",
            "Found existing installation: torchaudio 0.12.1+cu113\n",
            "Uninstalling torchaudio-0.12.1+cu113:\n",
            "  Successfully uninstalled torchaudio-0.12.1+cu113\n",
            "Found existing installation: torchtext 0.13.1\n",
            "Uninstalling torchtext-0.13.1:\n",
            "  Successfully uninstalled torchtext-0.13.1\n",
            "Found existing installation: fastai 2.7.9\n",
            "Uninstalling fastai-2.7.9:\n",
            "  Successfully uninstalled fastai-2.7.9\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install torch==1.10.2+cu102 -f https://download.pytorch.org/whl/torch_stable.html"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Gla-rF0jtQpO",
        "outputId": "abf871fc-fac9-4f74-f06e-d47f4f14ea71"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Looking in links: https://download.pytorch.org/whl/torch_stable.html\n",
            "Collecting torch==1.10.2+cu102\n",
            "  Downloading https://download.pytorch.org/whl/cu102/torch-1.10.2%2Bcu102-cp37-cp37m-linux_x86_64.whl (881.9 MB)\n",
            "\u001b[K     |██████████████████████████████▎ | 834.1 MB 1.2 MB/s eta 0:00:39tcmalloc: large alloc 1147494400 bytes == 0x39ad0000 @  0x7f9993f4c615 0x58e046 0x4f2e5e 0x4d19df 0x51b31c 0x5b41c5 0x58f49e 0x51b221 0x5b41c5 0x58f49e 0x51837f 0x4cfabb 0x517aa0 0x4cfabb 0x517aa0 0x4cfabb 0x517aa0 0x4ba70a 0x538136 0x590055 0x51b180 0x5b41c5 0x58f49e 0x51837f 0x5b41c5 0x58f49e 0x51740e 0x58f2a7 0x517947 0x5b41c5 0x58f49e\n",
            "\u001b[K     |████████████████████████████████| 881.9 MB 2.1 kB/s \n",
            "\u001b[?25hRequirement already satisfied: typing-extensions in /usr/local/lib/python3.7/dist-packages (from torch==1.10.2+cu102) (4.1.1)\n",
            "Installing collected packages: torch\n",
            "  Attempting uninstall: torch\n",
            "    Found existing installation: torch 1.12.1+cu113\n",
            "    Uninstalling torch-1.12.1+cu113:\n",
            "      Successfully uninstalled torch-1.12.1+cu113\n",
            "Successfully installed torch-1.10.2+cu102\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip freeze | grep torch"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2MB1FHZqtUuF",
        "outputId": "8ceee5ac-c068-4ccc-b234-913daa9eff8b"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "torch==1.10.2+cu102\n",
            "torchsummary==1.5.1\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Clone CGNN v1.0.2"
      ],
      "metadata": {
        "id": "ljCrFtDKtgqN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!git clone -b v1.0.2 https://github.com/Tony-Y/cgnn.git"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gfyk6EvLthfP",
        "outputId": "ff2b3f5b-f6e3-47fb-f51e-972fea7aae24"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Cloning into 'cgnn'...\n",
            "remote: Enumerating objects: 226, done.\u001b[K\n",
            "remote: Counting objects: 100% (90/90), done.\u001b[K\n",
            "remote: Compressing objects: 100% (67/67), done.\u001b[K\n",
            "remote: Total 226 (delta 38), reused 54 (delta 18), pack-reused 136\u001b[K\n",
            "Receiving objects: 100% (226/226), 2.04 MiB | 22.42 MiB/s, done.\n",
            "Resolving deltas: 100% (63/63), done.\n",
            "Note: checking out '15583437e0c1db175569bdb5523c718c93ce49cb'.\n",
            "\n",
            "You are in 'detached HEAD' state. You can look around, make experimental\n",
            "changes and commit them, and you can discard any commits you make in this\n",
            "state without impacting any branches by performing another checkout.\n",
            "\n",
            "If you want to create a new branch to retain commits you create, you may\n",
            "do so (now or later) by using -b with the checkout command again. Example:\n",
            "\n",
            "  git checkout -b <new-branch-name>\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Download the OQMD v1.2 Dataset"
      ],
      "metadata": {
        "id": "wtBG8jD7tnOk"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!wget -q \"https://zenodo.org/record/7118055/files/oqmd-v1_2-for-cgnn.7z?download=1\" -O oqmd.7z"
      ],
      "metadata": {
        "id": "pWTAa0dYtkdz"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!7z e oqmd.7z -ocgnn/OQMD *.npz *.json *.csv -r"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mZsfrxfWtrgK",
        "outputId": "14a2e45f-e20c-4ae0-8d20-2078b9bf18a8"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "7-Zip [64] 16.02 : Copyright (c) 1999-2016 Igor Pavlov : 2016-05-21\n",
            "p7zip Version 16.02 (locale=en_US.UTF-8,Utf16=on,HugeFiles=on,64 bits,2 CPUs Intel(R) Xeon(R) CPU @ 2.20GHz (406F0),ASM,AES-NI)\n",
            "\n",
            "Scanning the drive for archives:\n",
            "  0M Scan\b\b\b\b\b\b\b\b\b         \b\b\b\b\b\b\b\b\b1 file, 74747342 bytes (72 MiB)\n",
            "\n",
            "Extracting archive: oqmd.7z\n",
            "--\n",
            "Path = oqmd.7z\n",
            "Type = 7z\n",
            "Physical Size = 74747342\n",
            "Headers Size = 363\n",
            "Method = LZMA2:24\n",
            "Solid = +\n",
            "Blocks = 4\n",
            "\n",
            "  0%\b\b\b\b    \b\b\b\b  5% - oqmd_dataset/targets.csv\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                               \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 11% - oqmd_dataset/targets.csv\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                               \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 18% - oqmd_dataset/targets.csv\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                               \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 26% - oqmd_dataset/targets.csv\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                               \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 32% - oqmd_dataset/targets.csv\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                               \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 38% - oqmd_dataset/targets.csv\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                               \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 46% - oqmd_dataset/targets.csv\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                               \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 54% 2 - oqmd_dataset/split.json\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                                \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 57% 3 - oqmd_dataset/graph_data.npz\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                                    \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 61% 3 - oqmd_dataset/graph_data.npz\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                                    \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 65% 3 - oqmd_dataset/graph_data.npz\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                                    \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 69% 3 - oqmd_dataset/graph_data.npz\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                                    \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 73% 3 - oqmd_dataset/graph_data.npz\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                                    \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 76% 3 - oqmd_dataset/graph_data.npz\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                                    \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 80% 3 - oqmd_dataset/graph_data.npz\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                                    \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 84% 3 - oqmd_dataset/graph_data.npz\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                                    \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 89% 3 - oqmd_dataset/graph_data.npz\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                                    \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 93% 3 - oqmd_dataset/graph_data.npz\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                                    \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 95% 3 - oqmd_dataset/graph_data.npz\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                                    \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b 99% 3 - oqmd_dataset/graph_data.npz\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b                                    \b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\bEverything is Ok\n",
            "\n",
            "Files: 4\n",
            "Size:       143422415\n",
            "Compressed: 74747342\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!ls cgnn/OQMD/"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dkP3SRtDttxq",
        "outputId": "15c8c73d-5d49-4cd6-c2f6-650418a52201"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "config.json  graph_data.npz  oqmd.py  README.md  split.json  targets.csv\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Train a CGNN Model"
      ],
      "metadata": {
        "id": "QmiemtIatyw5"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "%pushd 'cgnn/examples/oqmd_fe_benchmark/'"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "06WFNzfbt3V5",
        "outputId": "a6324601-359a-4a87-c77d-c018411e0376"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content/cgnn/examples/oqmd_fe_benchmark\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['/content']"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!bash run_oqmd.sh | tee log.txt"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pZOC2STltzS0",
        "outputId": "0bf654b8-314d-4469-c680-9aea64d6d7ba"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Device: cuda\n",
            "\n",
            "Model: {'n_node_feat': 89, 'n_hidden_feat': 96, 'n_graph_feat': 192, 'n_conv': 4, 'n_fc': 2, 'activation': 'softplus', 'use_batch_norm': True, 'node_activation': None, 'use_node_batch_norm': False, 'edge_activation': None, 'use_edge_batch_norm': False, 'n_edge_net_feat': 16, 'n_edge_net_layers': 0, 'edge_net_activation': 'elu', 'use_edge_net_batch_norm': False, 'use_fast_edge_network': False, 'fast_edge_network_type': 0, 'use_aggregated_edge_network': False, 'edge_net_cardinality': 32, 'edge_net_width': 4, 'use_edge_net_shortcut': False, 'n_postconv_net_layers': 0, 'postconv_net_activation': 'elu', 'use_postconv_net_batch_norm': False, 'conv_bias': False, 'edge_net_bias': False, 'postconv_net_bias': False, 'full_pooling': True, 'gated_pooling': True, 'conv_type': 0, 'use_extension': False}\n",
            "\n",
            "Optimizer: {'optim': 'adam', 'lr': 0.001, 'weight_decay': 1e-06, 'clip_value': None}\n",
            "\n",
            "Scheduler: {'milestones': [250], 'gamma': 0.1, 'cosine_annealing': False}\n",
            "\n",
            "Dataset: {'dataset_path': '../../OQMD', 'target_name': 'formation_energy_per_atom', 'split_file': '../../OQMD/split.json'}\n",
            "\n",
            "Dataloader: {'num_workers': 0, 'batch_size': 512}\n",
            "\n",
            "Seed: 12345\n",
            "\n",
            "test: 55732 train: 449867 val: 56289\n",
            "Epoch 0/299\n",
            "current lr: 0.001\n",
            "train loss: 0.1727 mae: 0.2838\n",
            "val loss: 0.0848 mae: 0.2069\n",
            "Elapsed time (sec.): 16.710\n",
            "\n",
            "Epoch 1/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0726 mae: 0.1902\n",
            "val loss: 0.0648 mae: 0.1825\n",
            "Elapsed time (sec.): 14.405\n",
            "\n",
            "Epoch 2/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0537 mae: 0.1624\n",
            "val loss: 0.0430 mae: 0.1422\n",
            "Elapsed time (sec.): 14.358\n",
            "\n",
            "Epoch 3/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0440 mae: 0.1456\n",
            "val loss: 0.0438 mae: 0.1496\n",
            "Elapsed time (sec.): 16.352\n",
            "\n",
            "Epoch 4/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0382 mae: 0.1350\n",
            "val loss: 0.0366 mae: 0.1356\n",
            "Elapsed time (sec.): 14.285\n",
            "\n",
            "Epoch 5/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0344 mae: 0.1279\n",
            "val loss: 0.0327 mae: 0.1255\n",
            "Elapsed time (sec.): 16.326\n",
            "\n",
            "Epoch 6/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0313 mae: 0.1211\n",
            "val loss: 0.0234 mae: 0.1012\n",
            "Elapsed time (sec.): 14.167\n",
            "\n",
            "Epoch 7/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0282 mae: 0.1143\n",
            "val loss: 0.0352 mae: 0.1225\n",
            "Elapsed time (sec.): 14.067\n",
            "\n",
            "Epoch 8/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0270 mae: 0.1117\n",
            "val loss: 0.0223 mae: 0.0988\n",
            "Elapsed time (sec.): 16.226\n",
            "\n",
            "Epoch 9/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0259 mae: 0.1097\n",
            "val loss: 0.0205 mae: 0.0945\n",
            "Elapsed time (sec.): 14.217\n",
            "\n",
            "Epoch 10/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0239 mae: 0.1048\n",
            "val loss: 0.0223 mae: 0.1037\n",
            "Elapsed time (sec.): 16.265\n",
            "\n",
            "Epoch 11/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0227 mae: 0.1017\n",
            "val loss: 0.0185 mae: 0.0876\n",
            "Elapsed time (sec.): 14.151\n",
            "\n",
            "Epoch 12/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0220 mae: 0.1004\n",
            "val loss: 0.0190 mae: 0.0943\n",
            "Elapsed time (sec.): 14.007\n",
            "\n",
            "Epoch 13/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0207 mae: 0.0967\n",
            "val loss: 0.0171 mae: 0.0859\n",
            "Elapsed time (sec.): 16.141\n",
            "\n",
            "Epoch 14/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0204 mae: 0.0962\n",
            "val loss: 0.0163 mae: 0.0804\n",
            "Elapsed time (sec.): 14.011\n",
            "\n",
            "Epoch 15/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0197 mae: 0.0947\n",
            "val loss: 0.0202 mae: 0.0990\n",
            "Elapsed time (sec.): 16.213\n",
            "\n",
            "Epoch 16/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0186 mae: 0.0917\n",
            "val loss: 0.0164 mae: 0.0804\n",
            "Elapsed time (sec.): 14.134\n",
            "\n",
            "Epoch 17/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0183 mae: 0.0907\n",
            "val loss: 0.0164 mae: 0.0824\n",
            "Elapsed time (sec.): 14.022\n",
            "\n",
            "Epoch 18/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0177 mae: 0.0893\n",
            "val loss: 0.0175 mae: 0.0895\n",
            "Elapsed time (sec.): 16.090\n",
            "\n",
            "Epoch 19/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0172 mae: 0.0882\n",
            "val loss: 0.0172 mae: 0.0926\n",
            "Elapsed time (sec.): 14.096\n",
            "\n",
            "Epoch 20/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0168 mae: 0.0874\n",
            "val loss: 0.0154 mae: 0.0819\n",
            "Elapsed time (sec.): 16.136\n",
            "\n",
            "Epoch 21/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0162 mae: 0.0850\n",
            "val loss: 0.0132 mae: 0.0705\n",
            "Elapsed time (sec.): 14.026\n",
            "\n",
            "Epoch 22/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0168 mae: 0.0869\n",
            "val loss: 0.0136 mae: 0.0733\n",
            "Elapsed time (sec.): 14.070\n",
            "\n",
            "Epoch 23/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0161 mae: 0.0850\n",
            "val loss: 0.0121 mae: 0.0663\n",
            "Elapsed time (sec.): 16.121\n",
            "\n",
            "Epoch 24/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0155 mae: 0.0839\n",
            "val loss: 0.0137 mae: 0.0766\n",
            "Elapsed time (sec.): 13.988\n",
            "\n",
            "Epoch 25/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0152 mae: 0.0829\n",
            "val loss: 0.0138 mae: 0.0743\n",
            "Elapsed time (sec.): 16.165\n",
            "\n",
            "Epoch 26/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0148 mae: 0.0814\n",
            "val loss: 0.0132 mae: 0.0701\n",
            "Elapsed time (sec.): 14.036\n",
            "\n",
            "Epoch 27/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0152 mae: 0.0825\n",
            "val loss: 0.0157 mae: 0.0866\n",
            "Elapsed time (sec.): 14.155\n",
            "\n",
            "Epoch 28/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0144 mae: 0.0797\n",
            "val loss: 0.0117 mae: 0.0658\n",
            "Elapsed time (sec.): 16.131\n",
            "\n",
            "Epoch 29/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0142 mae: 0.0800\n",
            "val loss: 0.0109 mae: 0.0630\n",
            "Elapsed time (sec.): 14.066\n",
            "\n",
            "Epoch 30/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0141 mae: 0.0785\n",
            "val loss: 0.0115 mae: 0.0661\n",
            "Elapsed time (sec.): 14.122\n",
            "\n",
            "Epoch 31/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0133 mae: 0.0766\n",
            "val loss: 0.0118 mae: 0.0654\n",
            "Elapsed time (sec.): 16.064\n",
            "\n",
            "Epoch 32/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0138 mae: 0.0793\n",
            "val loss: 0.0121 mae: 0.0703\n",
            "Elapsed time (sec.): 14.103\n",
            "\n",
            "Epoch 33/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0132 mae: 0.0762\n",
            "val loss: 0.0122 mae: 0.0697\n",
            "Elapsed time (sec.): 16.085\n",
            "\n",
            "Epoch 34/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0133 mae: 0.0777\n",
            "val loss: 0.0115 mae: 0.0663\n",
            "Elapsed time (sec.): 14.074\n",
            "\n",
            "Epoch 35/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0129 mae: 0.0754\n",
            "val loss: 0.0126 mae: 0.0733\n",
            "Elapsed time (sec.): 16.153\n",
            "\n",
            "Epoch 36/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0129 mae: 0.0758\n",
            "val loss: 0.0104 mae: 0.0618\n",
            "Elapsed time (sec.): 14.178\n",
            "\n",
            "Epoch 37/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0124 mae: 0.0739\n",
            "val loss: 0.0148 mae: 0.0858\n",
            "Elapsed time (sec.): 14.040\n",
            "\n",
            "Epoch 38/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0126 mae: 0.0749\n",
            "val loss: 0.0114 mae: 0.0681\n",
            "Elapsed time (sec.): 16.112\n",
            "\n",
            "Epoch 39/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0122 mae: 0.0730\n",
            "val loss: 0.0107 mae: 0.0649\n",
            "Elapsed time (sec.): 13.992\n",
            "\n",
            "Epoch 40/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0124 mae: 0.0737\n",
            "val loss: 0.0111 mae: 0.0652\n",
            "Elapsed time (sec.): 16.117\n",
            "\n",
            "Epoch 41/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0123 mae: 0.0739\n",
            "val loss: 0.0111 mae: 0.0651\n",
            "Elapsed time (sec.): 14.069\n",
            "\n",
            "Epoch 42/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0118 mae: 0.0720\n",
            "val loss: 0.0140 mae: 0.0805\n",
            "Elapsed time (sec.): 14.147\n",
            "\n",
            "Epoch 43/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0116 mae: 0.0711\n",
            "val loss: 0.0101 mae: 0.0602\n",
            "Elapsed time (sec.): 16.168\n",
            "\n",
            "Epoch 44/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0119 mae: 0.0728\n",
            "val loss: 0.0099 mae: 0.0581\n",
            "Elapsed time (sec.): 14.065\n",
            "\n",
            "Epoch 45/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0124 mae: 0.0743\n",
            "val loss: 0.0155 mae: 0.0905\n",
            "Elapsed time (sec.): 16.104\n",
            "\n",
            "Epoch 46/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0116 mae: 0.0713\n",
            "val loss: 0.0101 mae: 0.0597\n",
            "Elapsed time (sec.): 14.045\n",
            "\n",
            "Epoch 47/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0117 mae: 0.0719\n",
            "val loss: 0.0098 mae: 0.0602\n",
            "Elapsed time (sec.): 14.053\n",
            "\n",
            "Epoch 48/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0117 mae: 0.0715\n",
            "val loss: 0.0097 mae: 0.0591\n",
            "Elapsed time (sec.): 16.162\n",
            "\n",
            "Epoch 49/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0110 mae: 0.0689\n",
            "val loss: 0.0142 mae: 0.0822\n",
            "Elapsed time (sec.): 13.951\n",
            "\n",
            "Epoch 50/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0109 mae: 0.0689\n",
            "val loss: 0.0133 mae: 0.0831\n",
            "Elapsed time (sec.): 14.056\n",
            "\n",
            "Epoch 51/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0112 mae: 0.0699\n",
            "val loss: 0.0096 mae: 0.0580\n",
            "Elapsed time (sec.): 16.136\n",
            "\n",
            "Epoch 52/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0107 mae: 0.0677\n",
            "val loss: 0.0119 mae: 0.0689\n",
            "Elapsed time (sec.): 13.996\n",
            "\n",
            "Epoch 53/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0112 mae: 0.0701\n",
            "val loss: 0.0096 mae: 0.0584\n",
            "Elapsed time (sec.): 16.077\n",
            "\n",
            "Epoch 54/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0107 mae: 0.0682\n",
            "val loss: 0.0096 mae: 0.0558\n",
            "Elapsed time (sec.): 14.037\n",
            "\n",
            "Epoch 55/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0108 mae: 0.0691\n",
            "val loss: 0.0112 mae: 0.0706\n",
            "Elapsed time (sec.): 13.984\n",
            "\n",
            "Epoch 56/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0106 mae: 0.0681\n",
            "val loss: 0.0097 mae: 0.0591\n",
            "Elapsed time (sec.): 16.016\n",
            "\n",
            "Epoch 57/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0107 mae: 0.0687\n",
            "val loss: 0.0128 mae: 0.0786\n",
            "Elapsed time (sec.): 13.995\n",
            "\n",
            "Epoch 58/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0108 mae: 0.0681\n",
            "val loss: 0.0097 mae: 0.0581\n",
            "Elapsed time (sec.): 16.081\n",
            "\n",
            "Epoch 59/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0103 mae: 0.0668\n",
            "val loss: 0.0097 mae: 0.0586\n",
            "Elapsed time (sec.): 14.053\n",
            "\n",
            "Epoch 60/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0105 mae: 0.0677\n",
            "val loss: 0.0095 mae: 0.0573\n",
            "Elapsed time (sec.): 16.132\n",
            "\n",
            "Epoch 61/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0106 mae: 0.0677\n",
            "val loss: 0.0097 mae: 0.0578\n",
            "Elapsed time (sec.): 14.025\n",
            "\n",
            "Epoch 62/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0107 mae: 0.0676\n",
            "val loss: 0.0108 mae: 0.0664\n",
            "Elapsed time (sec.): 13.981\n",
            "\n",
            "Epoch 63/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0103 mae: 0.0673\n",
            "val loss: 0.0115 mae: 0.0728\n",
            "Elapsed time (sec.): 16.057\n",
            "\n",
            "Epoch 64/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0100 mae: 0.0658\n",
            "val loss: 0.0122 mae: 0.0769\n",
            "Elapsed time (sec.): 13.995\n",
            "\n",
            "Epoch 65/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0099 mae: 0.0653\n",
            "val loss: 0.0097 mae: 0.0569\n",
            "Elapsed time (sec.): 16.107\n",
            "\n",
            "Epoch 66/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0099 mae: 0.0649\n",
            "val loss: 0.0093 mae: 0.0559\n",
            "Elapsed time (sec.): 13.971\n",
            "\n",
            "Epoch 67/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0100 mae: 0.0655\n",
            "val loss: 0.0103 mae: 0.0594\n",
            "Elapsed time (sec.): 13.982\n",
            "\n",
            "Epoch 68/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0099 mae: 0.0654\n",
            "val loss: 0.0096 mae: 0.0559\n",
            "Elapsed time (sec.): 16.099\n",
            "\n",
            "Epoch 69/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0098 mae: 0.0650\n",
            "val loss: 0.0093 mae: 0.0561\n",
            "Elapsed time (sec.): 14.017\n",
            "\n",
            "Epoch 70/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0101 mae: 0.0665\n",
            "val loss: 0.0095 mae: 0.0570\n",
            "Elapsed time (sec.): 16.079\n",
            "\n",
            "Epoch 71/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0095 mae: 0.0644\n",
            "val loss: 0.0093 mae: 0.0573\n",
            "Elapsed time (sec.): 14.125\n",
            "\n",
            "Epoch 72/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0096 mae: 0.0634\n",
            "val loss: 0.0095 mae: 0.0567\n",
            "Elapsed time (sec.): 13.964\n",
            "\n",
            "Epoch 73/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0107 mae: 0.0676\n",
            "val loss: 0.0089 mae: 0.0544\n",
            "Elapsed time (sec.): 16.052\n",
            "\n",
            "Epoch 74/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0097 mae: 0.0643\n",
            "val loss: 0.0105 mae: 0.0684\n",
            "Elapsed time (sec.): 13.969\n",
            "\n",
            "Epoch 75/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0092 mae: 0.0624\n",
            "val loss: 0.0094 mae: 0.0588\n",
            "Elapsed time (sec.): 16.153\n",
            "\n",
            "Epoch 76/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0092 mae: 0.0625\n",
            "val loss: 0.0093 mae: 0.0566\n",
            "Elapsed time (sec.): 14.068\n",
            "\n",
            "Epoch 77/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0093 mae: 0.0631\n",
            "val loss: 0.0093 mae: 0.0571\n",
            "Elapsed time (sec.): 14.023\n",
            "\n",
            "Epoch 78/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0097 mae: 0.0645\n",
            "val loss: 0.0109 mae: 0.0621\n",
            "Elapsed time (sec.): 16.016\n",
            "\n",
            "Epoch 79/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0093 mae: 0.0623\n",
            "val loss: 0.0092 mae: 0.0571\n",
            "Elapsed time (sec.): 13.963\n",
            "\n",
            "Epoch 80/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0093 mae: 0.0630\n",
            "val loss: 0.0104 mae: 0.0645\n",
            "Elapsed time (sec.): 16.168\n",
            "\n",
            "Epoch 81/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0090 mae: 0.0619\n",
            "val loss: 0.0089 mae: 0.0556\n",
            "Elapsed time (sec.): 14.054\n",
            "\n",
            "Epoch 82/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0092 mae: 0.0624\n",
            "val loss: 0.0100 mae: 0.0594\n",
            "Elapsed time (sec.): 13.868\n",
            "\n",
            "Epoch 83/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0100 mae: 0.0651\n",
            "val loss: 0.0090 mae: 0.0541\n",
            "Elapsed time (sec.): 16.068\n",
            "\n",
            "Epoch 84/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0090 mae: 0.0616\n",
            "val loss: 0.0101 mae: 0.0589\n",
            "Elapsed time (sec.): 13.945\n",
            "\n",
            "Epoch 85/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0093 mae: 0.0624\n",
            "val loss: 0.0093 mae: 0.0557\n",
            "Elapsed time (sec.): 16.135\n",
            "\n",
            "Epoch 86/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0088 mae: 0.0611\n",
            "val loss: 0.0090 mae: 0.0548\n",
            "Elapsed time (sec.): 13.981\n",
            "\n",
            "Epoch 87/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0087 mae: 0.0606\n",
            "val loss: 0.0097 mae: 0.0630\n",
            "Elapsed time (sec.): 14.010\n",
            "\n",
            "Epoch 88/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0090 mae: 0.0618\n",
            "val loss: 0.0091 mae: 0.0552\n",
            "Elapsed time (sec.): 16.065\n",
            "\n",
            "Epoch 89/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0087 mae: 0.0604\n",
            "val loss: 0.0092 mae: 0.0583\n",
            "Elapsed time (sec.): 13.958\n",
            "\n",
            "Epoch 90/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0093 mae: 0.0626\n",
            "val loss: 0.0113 mae: 0.0691\n",
            "Elapsed time (sec.): 13.931\n",
            "\n",
            "Epoch 91/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0089 mae: 0.0617\n",
            "val loss: 0.0103 mae: 0.0567\n",
            "Elapsed time (sec.): 16.081\n",
            "\n",
            "Epoch 92/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0090 mae: 0.0615\n",
            "val loss: 0.0103 mae: 0.0667\n",
            "Elapsed time (sec.): 14.061\n",
            "\n",
            "Epoch 93/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0090 mae: 0.0622\n",
            "val loss: 0.0092 mae: 0.0554\n",
            "Elapsed time (sec.): 16.126\n",
            "\n",
            "Epoch 94/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0093 mae: 0.0632\n",
            "val loss: 0.0090 mae: 0.0549\n",
            "Elapsed time (sec.): 13.933\n",
            "\n",
            "Epoch 95/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0088 mae: 0.0611\n",
            "val loss: 0.0087 mae: 0.0531\n",
            "Elapsed time (sec.): 16.098\n",
            "\n",
            "Epoch 96/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0085 mae: 0.0599\n",
            "val loss: 0.0106 mae: 0.0697\n",
            "Elapsed time (sec.): 14.056\n",
            "\n",
            "Epoch 97/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0088 mae: 0.0615\n",
            "val loss: 0.0095 mae: 0.0577\n",
            "Elapsed time (sec.): 14.018\n",
            "\n",
            "Epoch 98/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0087 mae: 0.0606\n",
            "val loss: 0.0094 mae: 0.0587\n",
            "Elapsed time (sec.): 16.006\n",
            "\n",
            "Epoch 99/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0088 mae: 0.0610\n",
            "val loss: 0.0106 mae: 0.0628\n",
            "Elapsed time (sec.): 14.366\n",
            "\n",
            "Epoch 100/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0086 mae: 0.0606\n",
            "val loss: 0.0098 mae: 0.0595\n",
            "Elapsed time (sec.): 16.257\n",
            "\n",
            "Epoch 101/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0086 mae: 0.0603\n",
            "val loss: 0.0094 mae: 0.0578\n",
            "Elapsed time (sec.): 14.400\n",
            "\n",
            "Epoch 102/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0087 mae: 0.0609\n",
            "val loss: 0.0098 mae: 0.0572\n",
            "Elapsed time (sec.): 14.263\n",
            "\n",
            "Epoch 103/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0089 mae: 0.0608\n",
            "val loss: 0.0085 mae: 0.0521\n",
            "Elapsed time (sec.): 16.072\n",
            "\n",
            "Epoch 104/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0085 mae: 0.0597\n",
            "val loss: 0.0091 mae: 0.0576\n",
            "Elapsed time (sec.): 13.947\n",
            "\n",
            "Epoch 105/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0089 mae: 0.0613\n",
            "val loss: 0.0086 mae: 0.0525\n",
            "Elapsed time (sec.): 16.113\n",
            "\n",
            "Epoch 106/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0085 mae: 0.0603\n",
            "val loss: 0.0087 mae: 0.0538\n",
            "Elapsed time (sec.): 13.885\n",
            "\n",
            "Epoch 107/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0083 mae: 0.0597\n",
            "val loss: 0.0094 mae: 0.0576\n",
            "Elapsed time (sec.): 13.958\n",
            "\n",
            "Epoch 108/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0084 mae: 0.0596\n",
            "val loss: 0.0089 mae: 0.0552\n",
            "Elapsed time (sec.): 16.137\n",
            "\n",
            "Epoch 109/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0084 mae: 0.0589\n",
            "val loss: 0.0089 mae: 0.0535\n",
            "Elapsed time (sec.): 13.934\n",
            "\n",
            "Epoch 110/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0082 mae: 0.0586\n",
            "val loss: 0.0114 mae: 0.0710\n",
            "Elapsed time (sec.): 13.976\n",
            "\n",
            "Epoch 111/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0084 mae: 0.0599\n",
            "val loss: 0.0088 mae: 0.0529\n",
            "Elapsed time (sec.): 15.943\n",
            "\n",
            "Epoch 112/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0086 mae: 0.0602\n",
            "val loss: 0.0090 mae: 0.0545\n",
            "Elapsed time (sec.): 13.922\n",
            "\n",
            "Epoch 113/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0085 mae: 0.0597\n",
            "val loss: 0.0092 mae: 0.0590\n",
            "Elapsed time (sec.): 16.154\n",
            "\n",
            "Epoch 114/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0082 mae: 0.0591\n",
            "val loss: 0.0087 mae: 0.0525\n",
            "Elapsed time (sec.): 13.969\n",
            "\n",
            "Epoch 115/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0084 mae: 0.0599\n",
            "val loss: 0.0088 mae: 0.0544\n",
            "Elapsed time (sec.): 13.965\n",
            "\n",
            "Epoch 116/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0082 mae: 0.0591\n",
            "val loss: 0.0086 mae: 0.0518\n",
            "Elapsed time (sec.): 15.965\n",
            "\n",
            "Epoch 117/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0091 mae: 0.0618\n",
            "val loss: 0.0090 mae: 0.0554\n",
            "Elapsed time (sec.): 13.980\n",
            "\n",
            "Epoch 118/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0081 mae: 0.0584\n",
            "val loss: 0.0083 mae: 0.0512\n",
            "Elapsed time (sec.): 16.016\n",
            "\n",
            "Epoch 119/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0081 mae: 0.0583\n",
            "val loss: 0.0094 mae: 0.0563\n",
            "Elapsed time (sec.): 14.028\n",
            "\n",
            "Epoch 120/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0083 mae: 0.0591\n",
            "val loss: 0.0091 mae: 0.0565\n",
            "Elapsed time (sec.): 16.037\n",
            "\n",
            "Epoch 121/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0084 mae: 0.0592\n",
            "val loss: 0.0116 mae: 0.0673\n",
            "Elapsed time (sec.): 14.010\n",
            "\n",
            "Epoch 122/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0089 mae: 0.0613\n",
            "val loss: 0.0088 mae: 0.0540\n",
            "Elapsed time (sec.): 13.928\n",
            "\n",
            "Epoch 123/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0079 mae: 0.0578\n",
            "val loss: 0.0104 mae: 0.0664\n",
            "Elapsed time (sec.): 16.018\n",
            "\n",
            "Epoch 124/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0079 mae: 0.0575\n",
            "val loss: 0.0104 mae: 0.0675\n",
            "Elapsed time (sec.): 13.935\n",
            "\n",
            "Epoch 125/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0079 mae: 0.0581\n",
            "val loss: 0.0087 mae: 0.0532\n",
            "Elapsed time (sec.): 16.083\n",
            "\n",
            "Epoch 126/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0080 mae: 0.0581\n",
            "val loss: 0.0121 mae: 0.0789\n",
            "Elapsed time (sec.): 13.889\n",
            "\n",
            "Epoch 127/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0080 mae: 0.0582\n",
            "val loss: 0.0088 mae: 0.0527\n",
            "Elapsed time (sec.): 13.902\n",
            "\n",
            "Epoch 128/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0081 mae: 0.0589\n",
            "val loss: 0.0091 mae: 0.0560\n",
            "Elapsed time (sec.): 15.998\n",
            "\n",
            "Epoch 129/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0083 mae: 0.0590\n",
            "val loss: 0.0101 mae: 0.0618\n",
            "Elapsed time (sec.): 13.945\n",
            "\n",
            "Epoch 130/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0083 mae: 0.0591\n",
            "val loss: 0.0090 mae: 0.0552\n",
            "Elapsed time (sec.): 16.117\n",
            "\n",
            "Epoch 131/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0079 mae: 0.0575\n",
            "val loss: 0.0087 mae: 0.0517\n",
            "Elapsed time (sec.): 14.030\n",
            "\n",
            "Epoch 132/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0079 mae: 0.0580\n",
            "val loss: 0.0100 mae: 0.0626\n",
            "Elapsed time (sec.): 13.965\n",
            "\n",
            "Epoch 133/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0080 mae: 0.0582\n",
            "val loss: 0.0091 mae: 0.0544\n",
            "Elapsed time (sec.): 16.056\n",
            "\n",
            "Epoch 134/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0081 mae: 0.0585\n",
            "val loss: 0.0097 mae: 0.0538\n",
            "Elapsed time (sec.): 14.041\n",
            "\n",
            "Epoch 135/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0081 mae: 0.0579\n",
            "val loss: 0.0093 mae: 0.0576\n",
            "Elapsed time (sec.): 16.005\n",
            "\n",
            "Epoch 136/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0078 mae: 0.0575\n",
            "val loss: 0.0093 mae: 0.0569\n",
            "Elapsed time (sec.): 14.000\n",
            "\n",
            "Epoch 137/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0101 mae: 0.0646\n",
            "val loss: 0.0102 mae: 0.0630\n",
            "Elapsed time (sec.): 13.890\n",
            "\n",
            "Epoch 138/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0084 mae: 0.0593\n",
            "val loss: 0.0096 mae: 0.0584\n",
            "Elapsed time (sec.): 15.986\n",
            "\n",
            "Epoch 139/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0078 mae: 0.0578\n",
            "val loss: 0.0087 mae: 0.0539\n",
            "Elapsed time (sec.): 13.854\n",
            "\n",
            "Epoch 140/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0077 mae: 0.0567\n",
            "val loss: 0.0109 mae: 0.0690\n",
            "Elapsed time (sec.): 16.068\n",
            "\n",
            "Epoch 141/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0080 mae: 0.0584\n",
            "val loss: 0.0086 mae: 0.0514\n",
            "Elapsed time (sec.): 13.966\n",
            "\n",
            "Epoch 142/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0078 mae: 0.0577\n",
            "val loss: 0.0093 mae: 0.0556\n",
            "Elapsed time (sec.): 13.886\n",
            "\n",
            "Epoch 143/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0077 mae: 0.0567\n",
            "val loss: 0.0085 mae: 0.0516\n",
            "Elapsed time (sec.): 15.956\n",
            "\n",
            "Epoch 144/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0079 mae: 0.0577\n",
            "val loss: 0.0093 mae: 0.0578\n",
            "Elapsed time (sec.): 13.930\n",
            "\n",
            "Epoch 145/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0079 mae: 0.0572\n",
            "val loss: 0.0099 mae: 0.0602\n",
            "Elapsed time (sec.): 16.035\n",
            "\n",
            "Epoch 146/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0084 mae: 0.0592\n",
            "val loss: 0.0086 mae: 0.0521\n",
            "Elapsed time (sec.): 13.925\n",
            "\n",
            "Epoch 147/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0077 mae: 0.0568\n",
            "val loss: 0.0085 mae: 0.0526\n",
            "Elapsed time (sec.): 13.933\n",
            "\n",
            "Epoch 148/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0076 mae: 0.0560\n",
            "val loss: 0.0086 mae: 0.0520\n",
            "Elapsed time (sec.): 15.981\n",
            "\n",
            "Epoch 149/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0095 mae: 0.0630\n",
            "val loss: 0.0099 mae: 0.0602\n",
            "Elapsed time (sec.): 13.831\n",
            "\n",
            "Epoch 150/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0081 mae: 0.0587\n",
            "val loss: 0.0086 mae: 0.0533\n",
            "Elapsed time (sec.): 13.938\n",
            "\n",
            "Epoch 151/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0076 mae: 0.0569\n",
            "val loss: 0.0099 mae: 0.0620\n",
            "Elapsed time (sec.): 15.955\n",
            "\n",
            "Epoch 152/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0078 mae: 0.0573\n",
            "val loss: 0.0088 mae: 0.0534\n",
            "Elapsed time (sec.): 13.910\n",
            "\n",
            "Epoch 153/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0078 mae: 0.0574\n",
            "val loss: 0.0091 mae: 0.0535\n",
            "Elapsed time (sec.): 16.004\n",
            "\n",
            "Epoch 154/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0077 mae: 0.0571\n",
            "val loss: 0.0092 mae: 0.0580\n",
            "Elapsed time (sec.): 13.909\n",
            "\n",
            "Epoch 155/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0074 mae: 0.0563\n",
            "val loss: 0.0094 mae: 0.0594\n",
            "Elapsed time (sec.): 16.113\n",
            "\n",
            "Epoch 156/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0077 mae: 0.0573\n",
            "val loss: 0.0093 mae: 0.0566\n",
            "Elapsed time (sec.): 13.958\n",
            "\n",
            "Epoch 157/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0077 mae: 0.0568\n",
            "val loss: 0.0098 mae: 0.0607\n",
            "Elapsed time (sec.): 13.937\n",
            "\n",
            "Epoch 158/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0081 mae: 0.0585\n",
            "val loss: 0.0098 mae: 0.0601\n",
            "Elapsed time (sec.): 15.998\n",
            "\n",
            "Epoch 159/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0079 mae: 0.0580\n",
            "val loss: 0.0092 mae: 0.0558\n",
            "Elapsed time (sec.): 13.877\n",
            "\n",
            "Epoch 160/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0075 mae: 0.0565\n",
            "val loss: 0.0085 mae: 0.0510\n",
            "Elapsed time (sec.): 16.030\n",
            "\n",
            "Epoch 161/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0082 mae: 0.0577\n",
            "val loss: 0.0106 mae: 0.0646\n",
            "Elapsed time (sec.): 14.009\n",
            "\n",
            "Epoch 162/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0076 mae: 0.0569\n",
            "val loss: 0.0090 mae: 0.0548\n",
            "Elapsed time (sec.): 13.944\n",
            "\n",
            "Epoch 163/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0078 mae: 0.0583\n",
            "val loss: 0.0083 mae: 0.0512\n",
            "Elapsed time (sec.): 16.007\n",
            "\n",
            "Epoch 164/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0079 mae: 0.0573\n",
            "val loss: 0.0089 mae: 0.0530\n",
            "Elapsed time (sec.): 13.946\n",
            "\n",
            "Epoch 165/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0076 mae: 0.0563\n",
            "val loss: 0.0090 mae: 0.0543\n",
            "Elapsed time (sec.): 16.047\n",
            "\n",
            "Epoch 166/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0077 mae: 0.0575\n",
            "val loss: 0.0109 mae: 0.0677\n",
            "Elapsed time (sec.): 14.005\n",
            "\n",
            "Epoch 167/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0077 mae: 0.0576\n",
            "val loss: 0.0087 mae: 0.0538\n",
            "Elapsed time (sec.): 13.903\n",
            "\n",
            "Epoch 168/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0077 mae: 0.0574\n",
            "val loss: 0.0089 mae: 0.0535\n",
            "Elapsed time (sec.): 16.002\n",
            "\n",
            "Epoch 169/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0075 mae: 0.0565\n",
            "val loss: 0.0087 mae: 0.0532\n",
            "Elapsed time (sec.): 13.927\n",
            "\n",
            "Epoch 170/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0076 mae: 0.0567\n",
            "val loss: 0.0092 mae: 0.0567\n",
            "Elapsed time (sec.): 13.890\n",
            "\n",
            "Epoch 171/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0074 mae: 0.0564\n",
            "val loss: 0.0097 mae: 0.0624\n",
            "Elapsed time (sec.): 15.847\n",
            "\n",
            "Epoch 172/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0076 mae: 0.0570\n",
            "val loss: 0.0089 mae: 0.0545\n",
            "Elapsed time (sec.): 13.862\n",
            "\n",
            "Epoch 173/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0075 mae: 0.0561\n",
            "val loss: 0.0093 mae: 0.0547\n",
            "Elapsed time (sec.): 16.021\n",
            "\n",
            "Epoch 174/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0080 mae: 0.0583\n",
            "val loss: 0.0096 mae: 0.0602\n",
            "Elapsed time (sec.): 13.987\n",
            "\n",
            "Epoch 175/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0077 mae: 0.0578\n",
            "val loss: 0.0100 mae: 0.0624\n",
            "Elapsed time (sec.): 13.890\n",
            "\n",
            "Epoch 176/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0073 mae: 0.0557\n",
            "val loss: 0.0089 mae: 0.0547\n",
            "Elapsed time (sec.): 16.028\n",
            "\n",
            "Epoch 177/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0078 mae: 0.0578\n",
            "val loss: 0.0088 mae: 0.0531\n",
            "Elapsed time (sec.): 13.919\n",
            "\n",
            "Epoch 178/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0075 mae: 0.0562\n",
            "val loss: 0.0091 mae: 0.0572\n",
            "Elapsed time (sec.): 15.978\n",
            "\n",
            "Epoch 179/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0074 mae: 0.0563\n",
            "val loss: 0.0083 mae: 0.0506\n",
            "Elapsed time (sec.): 13.908\n",
            "\n",
            "Epoch 180/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0075 mae: 0.0571\n",
            "val loss: 0.0093 mae: 0.0563\n",
            "Elapsed time (sec.): 16.031\n",
            "\n",
            "Epoch 181/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0076 mae: 0.0564\n",
            "val loss: 0.0087 mae: 0.0535\n",
            "Elapsed time (sec.): 13.967\n",
            "\n",
            "Epoch 182/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0076 mae: 0.0573\n",
            "val loss: 0.0085 mae: 0.0511\n",
            "Elapsed time (sec.): 13.934\n",
            "\n",
            "Epoch 183/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0073 mae: 0.0559\n",
            "val loss: 0.0085 mae: 0.0516\n",
            "Elapsed time (sec.): 15.900\n",
            "\n",
            "Epoch 184/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0074 mae: 0.0562\n",
            "val loss: 0.0087 mae: 0.0525\n",
            "Elapsed time (sec.): 13.923\n",
            "\n",
            "Epoch 185/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0074 mae: 0.0562\n",
            "val loss: 0.0093 mae: 0.0569\n",
            "Elapsed time (sec.): 16.027\n",
            "\n",
            "Epoch 186/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0073 mae: 0.0555\n",
            "val loss: 0.0088 mae: 0.0532\n",
            "Elapsed time (sec.): 13.911\n",
            "\n",
            "Epoch 187/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0073 mae: 0.0556\n",
            "val loss: 0.0088 mae: 0.0531\n",
            "Elapsed time (sec.): 13.943\n",
            "\n",
            "Epoch 188/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0075 mae: 0.0566\n",
            "val loss: 0.0090 mae: 0.0541\n",
            "Elapsed time (sec.): 15.941\n",
            "\n",
            "Epoch 189/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0074 mae: 0.0561\n",
            "val loss: 0.0087 mae: 0.0532\n",
            "Elapsed time (sec.): 13.933\n",
            "\n",
            "Epoch 190/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0074 mae: 0.0563\n",
            "val loss: 0.0100 mae: 0.0648\n",
            "Elapsed time (sec.): 16.034\n",
            "\n",
            "Epoch 191/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0072 mae: 0.0551\n",
            "val loss: 0.0090 mae: 0.0544\n",
            "Elapsed time (sec.): 13.972\n",
            "\n",
            "Epoch 192/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0074 mae: 0.0562\n",
            "val loss: 0.0112 mae: 0.0668\n",
            "Elapsed time (sec.): 13.851\n",
            "\n",
            "Epoch 193/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0075 mae: 0.0565\n",
            "val loss: 0.0086 mae: 0.0516\n",
            "Elapsed time (sec.): 15.951\n",
            "\n",
            "Epoch 194/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0072 mae: 0.0556\n",
            "val loss: 0.0088 mae: 0.0527\n",
            "Elapsed time (sec.): 13.915\n",
            "\n",
            "Epoch 195/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0070 mae: 0.0548\n",
            "val loss: 0.0093 mae: 0.0570\n",
            "Elapsed time (sec.): 16.108\n",
            "\n",
            "Epoch 196/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0072 mae: 0.0557\n",
            "val loss: 0.0114 mae: 0.0733\n",
            "Elapsed time (sec.): 14.000\n",
            "\n",
            "Epoch 197/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0075 mae: 0.0565\n",
            "val loss: 0.0106 mae: 0.0624\n",
            "Elapsed time (sec.): 13.993\n",
            "\n",
            "Epoch 198/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0082 mae: 0.0585\n",
            "val loss: 0.0092 mae: 0.0521\n",
            "Elapsed time (sec.): 15.964\n",
            "\n",
            "Epoch 199/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0072 mae: 0.0554\n",
            "val loss: 0.0083 mae: 0.0506\n",
            "Elapsed time (sec.): 13.903\n",
            "\n",
            "Epoch 200/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0071 mae: 0.0552\n",
            "val loss: 0.0099 mae: 0.0561\n",
            "Elapsed time (sec.): 15.911\n",
            "\n",
            "Epoch 201/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0071 mae: 0.0550\n",
            "val loss: 0.0100 mae: 0.0576\n",
            "Elapsed time (sec.): 13.976\n",
            "\n",
            "Epoch 202/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0070 mae: 0.0544\n",
            "val loss: 0.0086 mae: 0.0516\n",
            "Elapsed time (sec.): 13.914\n",
            "\n",
            "Epoch 203/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0080 mae: 0.0586\n",
            "val loss: 0.0086 mae: 0.0517\n",
            "Elapsed time (sec.): 16.001\n",
            "\n",
            "Epoch 204/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0074 mae: 0.0567\n",
            "val loss: 0.0091 mae: 0.0558\n",
            "Elapsed time (sec.): 13.868\n",
            "\n",
            "Epoch 205/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0069 mae: 0.0544\n",
            "val loss: 0.0086 mae: 0.0528\n",
            "Elapsed time (sec.): 15.971\n",
            "\n",
            "Epoch 206/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0072 mae: 0.0553\n",
            "val loss: 0.0086 mae: 0.0519\n",
            "Elapsed time (sec.): 13.927\n",
            "\n",
            "Epoch 207/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0072 mae: 0.0561\n",
            "val loss: 0.0094 mae: 0.0571\n",
            "Elapsed time (sec.): 13.918\n",
            "\n",
            "Epoch 208/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0070 mae: 0.0547\n",
            "val loss: 0.0087 mae: 0.0523\n",
            "Elapsed time (sec.): 15.893\n",
            "\n",
            "Epoch 209/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0072 mae: 0.0552\n",
            "val loss: 0.0091 mae: 0.0555\n",
            "Elapsed time (sec.): 13.929\n",
            "\n",
            "Epoch 210/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0075 mae: 0.0568\n",
            "val loss: 0.0084 mae: 0.0506\n",
            "Elapsed time (sec.): 13.925\n",
            "\n",
            "Epoch 211/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0070 mae: 0.0544\n",
            "val loss: 0.0083 mae: 0.0510\n",
            "Elapsed time (sec.): 16.014\n",
            "\n",
            "Epoch 212/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0071 mae: 0.0557\n",
            "val loss: 0.0092 mae: 0.0546\n",
            "Elapsed time (sec.): 13.917\n",
            "\n",
            "Epoch 213/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0072 mae: 0.0561\n",
            "val loss: 0.0092 mae: 0.0572\n",
            "Elapsed time (sec.): 16.037\n",
            "\n",
            "Epoch 214/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0071 mae: 0.0549\n",
            "val loss: 0.0084 mae: 0.0502\n",
            "Elapsed time (sec.): 13.865\n",
            "\n",
            "Epoch 215/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0072 mae: 0.0556\n",
            "val loss: 0.0084 mae: 0.0510\n",
            "Elapsed time (sec.): 16.006\n",
            "\n",
            "Epoch 216/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0070 mae: 0.0550\n",
            "val loss: 0.0091 mae: 0.0544\n",
            "Elapsed time (sec.): 13.969\n",
            "\n",
            "Epoch 217/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0071 mae: 0.0557\n",
            "val loss: 0.0088 mae: 0.0511\n",
            "Elapsed time (sec.): 13.942\n",
            "\n",
            "Epoch 218/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0072 mae: 0.0556\n",
            "val loss: 0.0084 mae: 0.0516\n",
            "Elapsed time (sec.): 16.140\n",
            "\n",
            "Epoch 219/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0071 mae: 0.0551\n",
            "val loss: 0.0083 mae: 0.0505\n",
            "Elapsed time (sec.): 13.926\n",
            "\n",
            "Epoch 220/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0072 mae: 0.0553\n",
            "val loss: 0.0099 mae: 0.0550\n",
            "Elapsed time (sec.): 15.911\n",
            "\n",
            "Epoch 221/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0072 mae: 0.0552\n",
            "val loss: 0.0117 mae: 0.0737\n",
            "Elapsed time (sec.): 13.941\n",
            "\n",
            "Epoch 222/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0071 mae: 0.0548\n",
            "val loss: 0.0087 mae: 0.0532\n",
            "Elapsed time (sec.): 13.919\n",
            "\n",
            "Epoch 223/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0069 mae: 0.0549\n",
            "val loss: 0.0095 mae: 0.0550\n",
            "Elapsed time (sec.): 16.007\n",
            "\n",
            "Epoch 224/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0071 mae: 0.0554\n",
            "val loss: 0.0090 mae: 0.0539\n",
            "Elapsed time (sec.): 13.862\n",
            "\n",
            "Epoch 225/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0068 mae: 0.0540\n",
            "val loss: 0.0087 mae: 0.0511\n",
            "Elapsed time (sec.): 15.936\n",
            "\n",
            "Epoch 226/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0070 mae: 0.0550\n",
            "val loss: 0.0089 mae: 0.0526\n",
            "Elapsed time (sec.): 13.965\n",
            "\n",
            "Epoch 227/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0068 mae: 0.0546\n",
            "val loss: 0.0085 mae: 0.0516\n",
            "Elapsed time (sec.): 13.871\n",
            "\n",
            "Epoch 228/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0069 mae: 0.0546\n",
            "val loss: 0.0091 mae: 0.0534\n",
            "Elapsed time (sec.): 16.018\n",
            "\n",
            "Epoch 229/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0070 mae: 0.0546\n",
            "val loss: 0.0086 mae: 0.0508\n",
            "Elapsed time (sec.): 13.925\n",
            "\n",
            "Epoch 230/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0077 mae: 0.0578\n",
            "val loss: 0.0084 mae: 0.0506\n",
            "Elapsed time (sec.): 13.898\n",
            "\n",
            "Epoch 231/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0073 mae: 0.0556\n",
            "val loss: 0.0087 mae: 0.0529\n",
            "Elapsed time (sec.): 15.947\n",
            "\n",
            "Epoch 232/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0075 mae: 0.0566\n",
            "val loss: 0.0091 mae: 0.0547\n",
            "Elapsed time (sec.): 13.880\n",
            "\n",
            "Epoch 233/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0070 mae: 0.0550\n",
            "val loss: 0.0086 mae: 0.0543\n",
            "Elapsed time (sec.): 15.969\n",
            "\n",
            "Epoch 234/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0065 mae: 0.0533\n",
            "val loss: 0.0083 mae: 0.0497\n",
            "Elapsed time (sec.): 13.961\n",
            "\n",
            "Epoch 235/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0067 mae: 0.0538\n",
            "val loss: 0.0093 mae: 0.0550\n",
            "Elapsed time (sec.): 13.886\n",
            "\n",
            "Epoch 236/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0067 mae: 0.0540\n",
            "val loss: 0.0094 mae: 0.0580\n",
            "Elapsed time (sec.): 15.927\n",
            "\n",
            "Epoch 237/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0069 mae: 0.0546\n",
            "val loss: 0.0088 mae: 0.0529\n",
            "Elapsed time (sec.): 13.877\n",
            "\n",
            "Epoch 238/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0070 mae: 0.0546\n",
            "val loss: 0.0104 mae: 0.0583\n",
            "Elapsed time (sec.): 16.019\n",
            "\n",
            "Epoch 239/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0071 mae: 0.0553\n",
            "val loss: 0.0099 mae: 0.0602\n",
            "Elapsed time (sec.): 13.966\n",
            "\n",
            "Epoch 240/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0068 mae: 0.0539\n",
            "val loss: 0.0086 mae: 0.0517\n",
            "Elapsed time (sec.): 16.066\n",
            "\n",
            "Epoch 241/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0068 mae: 0.0543\n",
            "val loss: 0.0085 mae: 0.0525\n",
            "Elapsed time (sec.): 13.902\n",
            "\n",
            "Epoch 242/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0067 mae: 0.0536\n",
            "val loss: 0.0091 mae: 0.0524\n",
            "Elapsed time (sec.): 13.880\n",
            "\n",
            "Epoch 243/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0070 mae: 0.0551\n",
            "val loss: 0.0093 mae: 0.0572\n",
            "Elapsed time (sec.): 15.936\n",
            "\n",
            "Epoch 244/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0066 mae: 0.0534\n",
            "val loss: 0.0102 mae: 0.0609\n",
            "Elapsed time (sec.): 13.925\n",
            "\n",
            "Epoch 245/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0067 mae: 0.0537\n",
            "val loss: 0.0087 mae: 0.0520\n",
            "Elapsed time (sec.): 16.007\n",
            "\n",
            "Epoch 246/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0082 mae: 0.0588\n",
            "val loss: 0.0095 mae: 0.0549\n",
            "Elapsed time (sec.): 13.953\n",
            "\n",
            "Epoch 247/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0069 mae: 0.0545\n",
            "val loss: 0.0095 mae: 0.0570\n",
            "Elapsed time (sec.): 13.909\n",
            "\n",
            "Epoch 248/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0067 mae: 0.0541\n",
            "val loss: 0.0098 mae: 0.0590\n",
            "Elapsed time (sec.): 15.950\n",
            "\n",
            "Epoch 249/299\n",
            "current lr: 0.001\n",
            "train loss: 0.0070 mae: 0.0545\n",
            "val loss: 0.0085 mae: 0.0508\n",
            "Elapsed time (sec.): 13.889\n",
            "\n",
            "Epoch 250/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0049 mae: 0.0445\n",
            "val loss: 0.0074 mae: 0.0440\n",
            "Elapsed time (sec.): 16.089\n",
            "\n",
            "Epoch 251/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0045 mae: 0.0429\n",
            "val loss: 0.0073 mae: 0.0430\n",
            "Elapsed time (sec.): 13.949\n",
            "\n",
            "Epoch 252/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0044 mae: 0.0425\n",
            "val loss: 0.0073 mae: 0.0429\n",
            "Elapsed time (sec.): 13.881\n",
            "\n",
            "Epoch 253/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0042 mae: 0.0420\n",
            "val loss: 0.0074 mae: 0.0432\n",
            "Elapsed time (sec.): 15.914\n",
            "\n",
            "Epoch 254/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0042 mae: 0.0420\n",
            "val loss: 0.0073 mae: 0.0429\n",
            "Elapsed time (sec.): 13.906\n",
            "\n",
            "Epoch 255/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0042 mae: 0.0421\n",
            "val loss: 0.0073 mae: 0.0425\n",
            "Elapsed time (sec.): 16.025\n",
            "\n",
            "Epoch 256/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0041 mae: 0.0415\n",
            "val loss: 0.0073 mae: 0.0428\n",
            "Elapsed time (sec.): 14.030\n",
            "\n",
            "Epoch 257/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0041 mae: 0.0414\n",
            "val loss: 0.0073 mae: 0.0426\n",
            "Elapsed time (sec.): 13.862\n",
            "\n",
            "Epoch 258/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0041 mae: 0.0416\n",
            "val loss: 0.0073 mae: 0.0428\n",
            "Elapsed time (sec.): 15.949\n",
            "\n",
            "Epoch 259/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0040 mae: 0.0411\n",
            "val loss: 0.0073 mae: 0.0425\n",
            "Elapsed time (sec.): 13.920\n",
            "\n",
            "Epoch 260/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0040 mae: 0.0414\n",
            "val loss: 0.0073 mae: 0.0424\n",
            "Elapsed time (sec.): 16.102\n",
            "\n",
            "Epoch 261/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0040 mae: 0.0410\n",
            "val loss: 0.0074 mae: 0.0437\n",
            "Elapsed time (sec.): 13.955\n",
            "\n",
            "Epoch 262/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0040 mae: 0.0414\n",
            "val loss: 0.0074 mae: 0.0423\n",
            "Elapsed time (sec.): 14.036\n",
            "\n",
            "Epoch 263/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0039 mae: 0.0407\n",
            "val loss: 0.0074 mae: 0.0427\n",
            "Elapsed time (sec.): 16.002\n",
            "\n",
            "Epoch 264/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0039 mae: 0.0406\n",
            "val loss: 0.0073 mae: 0.0423\n",
            "Elapsed time (sec.): 13.903\n",
            "\n",
            "Epoch 265/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0039 mae: 0.0406\n",
            "val loss: 0.0074 mae: 0.0429\n",
            "Elapsed time (sec.): 15.964\n",
            "\n",
            "Epoch 266/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0039 mae: 0.0408\n",
            "val loss: 0.0074 mae: 0.0435\n",
            "Elapsed time (sec.): 13.898\n",
            "\n",
            "Epoch 267/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0038 mae: 0.0404\n",
            "val loss: 0.0074 mae: 0.0425\n",
            "Elapsed time (sec.): 14.007\n",
            "\n",
            "Epoch 268/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0039 mae: 0.0410\n",
            "val loss: 0.0074 mae: 0.0428\n",
            "Elapsed time (sec.): 15.951\n",
            "\n",
            "Epoch 269/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0038 mae: 0.0405\n",
            "val loss: 0.0076 mae: 0.0436\n",
            "Elapsed time (sec.): 13.918\n",
            "\n",
            "Epoch 270/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0038 mae: 0.0405\n",
            "val loss: 0.0075 mae: 0.0431\n",
            "Elapsed time (sec.): 13.872\n",
            "\n",
            "Epoch 271/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0038 mae: 0.0404\n",
            "val loss: 0.0074 mae: 0.0425\n",
            "Elapsed time (sec.): 15.968\n",
            "\n",
            "Epoch 272/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0039 mae: 0.0410\n",
            "val loss: 0.0074 mae: 0.0430\n",
            "Elapsed time (sec.): 13.850\n",
            "\n",
            "Epoch 273/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0038 mae: 0.0403\n",
            "val loss: 0.0074 mae: 0.0424\n",
            "Elapsed time (sec.): 15.955\n",
            "\n",
            "Epoch 274/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0037 mae: 0.0400\n",
            "val loss: 0.0074 mae: 0.0423\n",
            "Elapsed time (sec.): 13.869\n",
            "\n",
            "Epoch 275/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0038 mae: 0.0404\n",
            "val loss: 0.0074 mae: 0.0421\n",
            "Elapsed time (sec.): 15.980\n",
            "\n",
            "Epoch 276/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0037 mae: 0.0402\n",
            "val loss: 0.0075 mae: 0.0425\n",
            "Elapsed time (sec.): 13.909\n",
            "\n",
            "Epoch 277/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0037 mae: 0.0401\n",
            "val loss: 0.0075 mae: 0.0428\n",
            "Elapsed time (sec.): 13.914\n",
            "\n",
            "Epoch 278/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0037 mae: 0.0402\n",
            "val loss: 0.0075 mae: 0.0426\n",
            "Elapsed time (sec.): 15.971\n",
            "\n",
            "Epoch 279/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0038 mae: 0.0405\n",
            "val loss: 0.0075 mae: 0.0428\n",
            "Elapsed time (sec.): 13.933\n",
            "\n",
            "Epoch 280/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0037 mae: 0.0401\n",
            "val loss: 0.0075 mae: 0.0424\n",
            "Elapsed time (sec.): 15.925\n",
            "\n",
            "Epoch 281/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0037 mae: 0.0400\n",
            "val loss: 0.0074 mae: 0.0423\n",
            "Elapsed time (sec.): 14.029\n",
            "\n",
            "Epoch 282/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0037 mae: 0.0398\n",
            "val loss: 0.0075 mae: 0.0431\n",
            "Elapsed time (sec.): 13.881\n",
            "\n",
            "Epoch 283/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0036 mae: 0.0396\n",
            "val loss: 0.0075 mae: 0.0431\n",
            "Elapsed time (sec.): 16.007\n",
            "\n",
            "Epoch 284/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0036 mae: 0.0393\n",
            "val loss: 0.0075 mae: 0.0425\n",
            "Elapsed time (sec.): 13.901\n",
            "\n",
            "Epoch 285/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0036 mae: 0.0398\n",
            "val loss: 0.0075 mae: 0.0423\n",
            "Elapsed time (sec.): 15.951\n",
            "\n",
            "Epoch 286/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0037 mae: 0.0398\n",
            "val loss: 0.0075 mae: 0.0428\n",
            "Elapsed time (sec.): 13.865\n",
            "\n",
            "Epoch 287/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0036 mae: 0.0394\n",
            "val loss: 0.0076 mae: 0.0431\n",
            "Elapsed time (sec.): 13.885\n",
            "\n",
            "Epoch 288/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0037 mae: 0.0399\n",
            "val loss: 0.0076 mae: 0.0425\n",
            "Elapsed time (sec.): 15.963\n",
            "\n",
            "Epoch 289/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0036 mae: 0.0396\n",
            "val loss: 0.0075 mae: 0.0423\n",
            "Elapsed time (sec.): 13.881\n",
            "\n",
            "Epoch 290/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0036 mae: 0.0395\n",
            "val loss: 0.0078 mae: 0.0448\n",
            "Elapsed time (sec.): 13.849\n",
            "\n",
            "Epoch 291/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0036 mae: 0.0400\n",
            "val loss: 0.0075 mae: 0.0421\n",
            "Elapsed time (sec.): 15.898\n",
            "\n",
            "Epoch 292/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0035 mae: 0.0390\n",
            "val loss: 0.0075 mae: 0.0421\n",
            "Elapsed time (sec.): 13.868\n",
            "\n",
            "Epoch 293/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0036 mae: 0.0395\n",
            "val loss: 0.0075 mae: 0.0428\n",
            "Elapsed time (sec.): 15.934\n",
            "\n",
            "Epoch 294/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0036 mae: 0.0399\n",
            "val loss: 0.0075 mae: 0.0420\n",
            "Elapsed time (sec.): 13.899\n",
            "\n",
            "Epoch 295/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0036 mae: 0.0395\n",
            "val loss: 0.0075 mae: 0.0422\n",
            "Elapsed time (sec.): 13.887\n",
            "\n",
            "Epoch 296/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0036 mae: 0.0397\n",
            "val loss: 0.0075 mae: 0.0427\n",
            "Elapsed time (sec.): 15.896\n",
            "\n",
            "Epoch 297/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0036 mae: 0.0396\n",
            "val loss: 0.0075 mae: 0.0420\n",
            "Elapsed time (sec.): 13.862\n",
            "\n",
            "Epoch 298/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0035 mae: 0.0394\n",
            "val loss: 0.0076 mae: 0.0428\n",
            "Elapsed time (sec.): 15.903\n",
            "\n",
            "Epoch 299/299\n",
            "current lr: 0.0001\n",
            "train loss: 0.0036 mae: 0.0395\n",
            "val loss: 0.0076 mae: 0.0428\n",
            "Elapsed time (sec.): 13.847\n",
            "\n",
            "Total elapsed time (sec.): 4440.877\n",
            "The best val metric: 0.041993\n",
            "\n",
            "test loss: 0.0074 mae: 0.0424\n",
            "\n",
            "END\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%run '../../tools/plot_history.py'"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 297
        },
        "id": "jdc-Ci-Mt8JH",
        "outputId": "4bddc7c0-006f-46af-e8d2-f72e85625930"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%run '../../tools/plot_test.py'"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 368
        },
        "id": "goAZM7emt-in",
        "outputId": "bea4e1d2-f049-4c53-e397-76121c39d813"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "RMSE: 0.08612316914923318\n",
            "MAE: 0.04243335051803046\n",
            "max_value 4.068542\n",
            "min_value -4.449743\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAEGCAYAAABsLkJ6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3de5zU5XX48c/5fufC/VJADCyIiJcCAaJrMEGNkqQxCWpT0XjLpYnSXIzmFyPGWLXWV5MoSWwabVNqbLShNSiNKGoSrSYGAsTVLAhozHrhmnDZIOzCMtfz+2NmlpnZmdmZ3Zn5zuyc9+vFS3auR4Xn+X7Pc57ziKpijDGm8TheB2CMMcYbNgEYY0yDsgnAGGMalE0AxhjToGwCMMaYBuXzOoBSjB07VqdMmeJ1GMYYU1defPHFfao6LvvxupoApkyZQktLi9dhGGNMXRGRrbketxSQMcY0KJsAjDGmQdkEYIwxDcomAGOMaVA2ARhjTIOyCcAYYxqUTQDGGFPD2jtDbNj+Nu2dobJ/tuf7AETEBVqAnaq6wOt4jDGmVqxs3cmNKzbidxwi8Th3XTSLC+ZMLNvn18IdwHXAK14HYYwxtaS9M8SNKzZyJBKnIxTlSCTO4hUby3on4OkEICJNwEeB+7yMwxhjas2O/V34ncwh2u847NjfVbbv8PoO4J+BxUA83wtEZJGItIhIy969e6sXmTHGeKhp9GAi8cyhMRKP0zR6cNm+w7MJQEQWAHtU9cVCr1PVpararKrN48b16GVkjDED0phhQe66aBaD/A7Dgz4G+R3uumgWY4YFy/YdXi4CzwMuEJGPAIOAESLyY1W90sOYjDGmZlwwZyLzpo1lx/4umkYPLuvgDx5OAKp6E3ATgIicA3zVBn9jjMk0Zliw7AN/itdrAMYYYzzi+T4AAFX9JfBLj8MwxpiGYncAxhhTIdm7eCu5q7cvauIOwBhjBor2zhA79nex/o12lvzi9/gcIa7Kx0+fxPKWHRXb1dsXNgEYY0yZpFo3qEIomqjhj8QUgAfXbgPgSHLb0+IVG5k3bWzFFniLYSkgY4wpg7bdHdzwSKJ1Q2rwL6Tcu3r7wu4AjDGmH9o7Qyxbv43vP/sakVjx7yv3rt6+sAnAGGP6aNm6rdz++GbCyTRPIa4IPhcCrtu9BuBl+gdsAjDGmD5Ztm4rNz+6qajX+l3hOxfPruiu3r6wCcAYY0rQ3hli7ev7uO2xzXlf4wr8ZNEZHI7EAWXGhJHdA34tDPwpNgEYY0wBqbLOptGDWd22j68+vKG7sicfxxGOHzespgb7XGwCMMaYPNJP5ArHYsSVXgd/gEE+lx37u2wCMMaYepR+IteR/EeW5FQLFT7FsH0AxhhDzzYNa1/fh/Z+sY/rwJffP42gTxjidwn6pCYqfIphdwDGmIaXSvW4IkRicd45cQQvbjtQ1Ht9ItzzXBsigusAKpUNtozsDsAY09DSUz2HwjHCMS168AcIxZRoPLE2cDgcIxQt/+HtlWITgDGmoe3Y31X2gbAW2jwUw7MUkIgMAp4Hgsk4HlHV27yKxxjTWNp2d7C6bS+v/LEjWa9fPvWyCOzlGkAImK+qnSLiB1aLyFOqus7DmIwxdS69bj/fQuziRzawvGVHSZ/rdwGEgM8hElPi8TjZPd+GBFziqrYI3BtVVaAz+aM/+auINXdjjMktvW4/X8/9pb96veTBHyASg69/+GTmTh3D0IDLgntWE40fnQGCPuEHV56aseu31nm6BiAiroi0AnuAp1V1vZfxGGPqV/pibkcoypFIz8XYljfb+cZTr/b5O777zGs0jR7MtPHDueuiWQzyOwwP+hjkd1iycDZnn3RM3Qz+4HEZqKrGgDkiMgr4qYjMVNWM7koisghYBDB58mQPojTG1IMd+7vwO07Gpq3UYuyYYUEWP9zK8hd39us70j/vgjkTa665W6lqogpIVd8GngPOy/HcUlVtVtXmcePGVT84Y0xdaBo9mEg8Myl/JBojEo3xuf9q6ffgDz0Xd8cMCzJ70qi6HPzBwwlARMYlr/wRkcHAB4G+35sZYxramGFB7rpoFr60US0SUxb++zp+tnl3vz57aNBlkN+pm8XdYnmZAnoH8ICIuCQmouWqusrDeIwxdW7etLGAUK56Ep8Dt18wk5kTR9ZtmqcQL6uANgLv8ur7jTH1pVB5Z3tniM27DrLujX1E4+UrJnQdh/NmHjvgBv4U6wVkjKl5hco7V7bu5PrlrT1q8ssh4Dp10da5r2piEdgYY/IpVN7Z3hli8SMb+jT4f+GcqXz9w6cQcIWhgUQXT7+b2citXnb09pXdARhjalquXj1+x2HzrgPsORgiFC095eNz4LNnTmXMsCAXndbUnVpa07aPxVl3GgP16h9sAjCmLhTT3mCgxrFs/dYevXo6Q1E+ef8LJX+W3xEcB5YsnJ1xRm/q9wOhtr8UNgEYU+OKaW8wUONo292Rs21DX5Z5v3DOVM6YOpYZE0YUHNjTJ4SBztYAjKlhxbQ3GGhxtO3u4Edr3mTVhp3826/ayva5D/xmK4v+q4U1bfvK9pn1zu4AjKlhvbU3KKSc6ZpccTgIm3cd5OyTyrdD/9ZHX+bBddvK9nnpDoVjACxesZF508Y2zFV+IXYHYEwNy9XeoJjKlJWtO5l357Nced965t35LI+19q8NQq44DkdiXP1gS78/O6Vtd0dZB/+gmzivN+jLrOypl8NaqsEmAGNqWKq9QXrXyd4qU8qdrkndSdyyYHqPwbQ/xx9mH8L+r798vU/x5XJJcxPLPzePn193NiKNVdpZCksBGVPjSq1M6U/aKCU16G/aeYA7ntjSvfD7mXnH88DarRxOplPyfXZv6af0BeVwLMZZJ47jmVf2FBVbMR7bsIsbzzulewLNLu0E2LD97Yao9CnEJgBj6kAplSl9TRulpAZnnyN0hhIDfWoyuX/NmyR67eT/7N6qhdLvUFKfW87BHwq3bV7dto95dz7reVVVLbAUkDEDTF/SRinpg3Nq8E8XcF2uOXda3s8uJv20Y38XWsZ+Pbl0RaI52zYDNVFVVSvsDsCYAaivG5pypY/SReJxLp87mcvnTs752bne74p0X423d4bYsusAoVj5JgCfQ49WENl5/0LxlZoeG0hsAjBmgOrLhqZc6SOAoQGXWNZh57k+O9f7D4VjbNp1gLfaD1WkaVuuqSTgyz2o9zc9NtDYBGDMANWXfQC5Fk1vWTCdmROK64c/ZliQr3zwJL7xZObZTrc/tom49rxS76+AK7iO0JU9qMc056Ceb1G4Ea/+wSYAYwaM9AF/ddu+Prdt6E8/nGXrtnLXz3oe7BfuuZxQFiKQaznhtvOn54270fr9FOLZBCAik4AHgfEk7uKWqur3vIrHmHqWWVYZJxaPE40frd4pdfdr6nWpDVOp/H2uQTP1+Lo32vnmU9U71TXoc1iyMFHSuXjFRlwRIrE4t50/gyvmHlfwvY3U76cQL+8AosD1qvqSiAwHXhSRp1V1i4cxGVN3cpVVZit1oTO7lPOS5iaWt+zAFQjFlMUfOpn5Jx/D/Wve4uGWbTgiZV3Y7Y3fgf/4ZHN3Gwq7ou8bL4+E/CPwx+TvO0TkFWAiYBOAMSXorXIHIByLcaArTHtnqNcBsm13Bzc8spFw9OiE8uDazBYN33jy1aw8f/UGfwDHcZgxYUT3z3ZF3zc1sQ9ARKaQOB94fY7nFolIi4i07N27t9qhGVPzclW2OFlVkJGY8sVlv+u1L9CydVs573vPE67E+YpldM2502zALwPPJwARGQasAL6sqgezn1fVpararKrN48aVr+ugMQNF9savoC9RGZMurmRsfGrb3cHzr+3l+df2dG+CWvqr17n50U0VOVu3nII+4fK5k70OY0DwtApIRPwkBv9lqvq/XsZiTD1Lr2w50BXmi8t+RyQWzflajSsf+ufnSaXs/a7wsXdNzHnwSq0J+oQlC2cDtdXLp1ZObCuVl1VAAvwQeEVVv+tVHMYMFKk8eHtnKOdmrpTsxdpITGtq8L/6zOO5f82bZK8pf/59U7nqrKk118unVk5s6wsvU0DzgE8A80WkNfnrIx7GY8yAkJ0S8ruCz4HhQR8BVwi4udskeM11hH/62ExuXjCduz8+h6DPYbDfJeAmHr/xw38J1FYvn1o5sa2vvKwCWk12W0FjTFlkb3aCRLVQJBrjsvt61Fp4yu8IV519PFedORVIpHbmTRvLE186k9btbzNn0iimjR8O1F4vn1qLp1S2E9iYASq7NDK1O9jzyo80fhd++KnTOfukcRmplCPRGKrKYL8vI61Sa718ai2eUtXSnwVjTIWkpyqquWGrN5EYPPvKbtp2d2SkUiIxJRqnR1qlP62uK6HW4imV3QEY0wCK2SzmlR+t3cp//3Y7UmAzWfYBL9PfMaJHesgr9dxbyCYAYzzS3hli866DgDJjwkiAkgeRYsoP2ztDHOgKE45VqCNbCXL17gcIxwpPTOlplVqsuqnXncg2ARjjgZWtOzN64zuSqIIZ5HPzHqOYPdAXMxCmvyauiQF4sN9HVyRa9Q1fn3/fVEYPCfCNPA3jBvkd4nEl6HNzrgGkSlyz+x6V2ujOHGUTgDFV1t4ZYvEjGzIG4LhCPKbdm7fSB7VcA/28aWN7DIQ3PLKB6e8YwaFwrPtqOfs1QZ/DvVecyoqXtrOy9Y9V/ff+0W/eIq5wSXMTP/3dTiI51iKevPasjPizJ716r7qpNTYBGFNlm3cdIFEB3XvOG3oO4otXbGTpJ07rMRCGookdvkMCPsKxOAtPbUI08zsCrsNvXt9X9cEfoCuSiPWnv9vJ3ZfMYfOuA9y/5i0C7tGJLTufbyd6VZZNAMZU0crWnSx+ZAOhaOFKnNSglu+KFyRnTj+W7PkDsOy323o8fzgS499/9Ub//iX6KRJTrl/eijjCrQumM3NicaeNgZ3oVW42ARhTJan8dW+DP8Dpx43uHtSyr3hD0Rhv7O3kY++ayEMvlNbCIZbr+Kwyu/T0Jla8lDvFkxKKKcSU2x/fzJPXnlXSAF7PVTe1xvYBGFMlqav5Yvy6rZ223R096swdgXBM+YfHt5Q8+FfLR975Dtbd9H4e/Mzp3HPZuyjUeSIcUz7y/dUFW1TnMmZYkNmTRtng3082ARhTJU2jB9MVyd2hM5fW7W8DiSveNTfO55t/MzPn+be1ZntyQfbsk45hwewJ3P3xOfgLzALhaH31zxlIbAIwpoLaO0Ns2P529+CWaIJbnCljhmT8/NruzrLGVil3rNqSMZhfMGciT117VsEmdOmL3qZ6bA3AmArJPKg9xsJTJyUqXrIWb/2ucMbxf8Gv29q7H/M5cOX9v+Wui2ahJCqBarSJZw+uSI+yzGnjh/Pti2ezeMVGHBEOhzP/G1gljzdsAjCmAnJtWMpVlQMgCBc3T+K695/IpfetJ5rsgxONx/nqw62IOIRq/ZiuNPnOH05fvN208wB3PLHFKnk8ZhOAMRVQSu+dcCzO9Q9v4IefOp3BPpeOtJO8wjHwu5mJfwc87ejjCvzV9GP42eY9OXcyxOPw+R+/REy1x+7kVMuE2ZNGcd7MY62Sx2OergGIyP0iskdENnkZhzHlNjTgEooW33snElMOdkVy1vZnl1N6NfinMlAxhafyDP6p5w+FY70ejmKVPN7zehH4R8B5HsdgTFmtbN3JgntW4zilJe1b3voz15x7YoWi6r++FCCl1gP6InsB3ZSfpykgVX1eRKZ4GYMx5ZSe+08pNmXzo7VbWXhqfZwlW6xIrG+Lu7XY8XMg8voOoFciskhEWkSkZe/evV6HY0xBuTZ7lZKyeeSl0jZE1YKhQZegT3LW+t92/oySUzyFztm1u4LyqvlFYFVdCiwFaG5uroNtMKZRJfruR3rtbT+QXDD7WD575gk0jR7MmrZ9LF6xEdcRIjHltvOnc8Xc40r+zHz9j5at38a//rLN7grKqOgJQEQmAselv0dVn69EUMbUm1TKwhUhUqBkM+hCyPtzWcrmF1v2cNv5M7tP6ipHj55cHT/DsRj3PtdGKGrnAJRTUSkgEbkTWAP8PXBD8tdXKxiXMXUjPWVxKBwrmPIJD6DBHxKH2KQv8pajsifXObvXnHsiATdzuLLdw/1X7B3AXwMnq2pZE28i8j/AOcBYEdkB3KaqPyzndxhTaTv2dxEvsknPQMthHgrF2LTzALMnjSrr52bfTQDc+8u2jNfY7uH+K3YR+A3AX+4vV9XLVPUdqupX1SYb/E0tS1+ATP/90IBLuEDr43ozd8rokl5/xxNbKrIom343keuuwHYP91+xdwCHgVYR+T+g+/+0ql5bkaiMqTHpZYnZ59X+7Xun4HelYP/7evLitv24ktjQVYxcvX8qwc4BKL9iJ4DHkr+MGTByHbSe6znoeSwjHD156988PmGr3EptOxSOxqqWikndDZjyKGoCUNUHRCQAnJR86PeqGqlcWMaUX/qgvrptX96NRtmbkL54zrSi+/oMVK7AJ844jh+t3drjuavOmmqDcp0qagIQkXOAB4C3SLQEmSQin7IyUFMvslszxzXRYye7pBB6Xu3f81wbA2/5tjSL3jeVq86cyrLfbstIdfld4aqzpuZ9X6G7LOO9YheBvwP8laq+T1XPBj4E3F25sIwpn+ydpaGo9sjXp0oKc+3kDbgOn5l3fDVD9ky+AeH+1W+y/1CYa+efSNDnMCTgEvQ5fOfi2XkH9pWtO5l357Nced965t35bMnHPprKK3YNwK+qv0/9oKqviUjZq4KMqYQd+7vQXso000sKszchZf88kMWB5skjadl2IONxEeEj//Jrgj4XUD7/vml8eOaxHArHevT9h9znIdjGrdpT7ATQIiL3AT9O/nwF0FKZkIwpr6EBl1COkpaAKwR9bo8DSe66aBaL09YAblkwnX98fHO1w/ZM9uAPdDe3CyfPKvjnZ17jnufaEiec5WjLkK+dQzWqhUzxip0APg98EUiVff4a+NeKRGRMmR0Kx3qUafpd4b5PNTNycKA7P922u4PVbXsZO2wQq645k0PhRHXL5l0HccUBBtg23n6IKcSi8e6TyrKv7nO1c+jPxi1bS6iMYquAQsB3k7+MqStDA26PnH8kpkwYOZhp44cDcOujL/PguqNHNgrwvUvn8Fb7IRY/srGujmT0QvbVfWrj1uKsSqu+DN7WGrpyCk4AIrJcVS8RkZfJUQahqrMqFpkxZXIoHCPoSkYaKOgKuw50cSgcIxKNZQz+kPjDft1DrQR89XUer1dyXd2XY+OWrSVUVm93ANcl/7mg0oEYUylNowcjTubW1jhw1QMv4HNcQjmOYYTEJGCDf2FDgy6xuOa9uu/vxi1bS6isghOAqv4x+dsvqOqN6c8lO4Te2PNdxtSW7HREOBYnHI2jkPMM3kaTSJHFcRzJOMms0Otjqtzy0enMnDiy153UxQzU+V5f7rUEk6nYReAP0nOw/3COx4ypSenpiO1/Psw1//M7r0OqmoALqpA9tvscuP3CmcycMJKhAZcF96zOeN7vJNo9H4kevXMaGnS5/fwZnHvKMQXr/0vJ2Rd6fTnXEkxPva0BfB74AnCCiGxMe2o48JtKBmZMuaXSEQe6wl6HUjWffM9kmkYN4RtPvdrjOddxOG/GsfnLXz86nTue2EL68l8srgUH/1Jz9m27O7jhkY2ECxz0Yk3gKqe3O4D/Bp4Cvgl8Le3xDlX9c8WiMqaCZkwYOaC6d+bjd+DkY0dw+2O59zCkDnMpNNAOH+Qr6eq7lJz9ytad3PDwhh6ttHO93prAVUZvawAHgAMi8j3gz6raASAiI0Rkrqqu78+Xi8h5wPcAF7hPVb/Vn88zphhjhgW57N2TeHDttt5fXMcicbj98S3kOKsdSFzNZ+fSswfaXJNCofx+sTn71J1CrnMULMdfPcX2Avo3oDPt587kY30mIi5wL4m1hOnAZSIyvT+faRpT+uEsxb5+ecuOCkdVG3wOeQ+r+cy8KUVdVacfzNJbf59iD27J1XMJIOCzg16qqdhFYFHV7j9FqhoXkaIPlM/j3UCbqr4BICIPARcCW/r5uaaB9GWT0OZdB4gVeYRjvTscjnP2iWN5/g/7ejz3nhPGlvRZxeb3i8nZ57pTCLjCk186s3tznqm8oo+EFJFrRcSf/HUdiWMi+2MisD3t5x3JxzKIyCIRaRGRlr179/bzK81Akt3l80gkzuIVG2l5s51HWrbTtrujx3tWtu7ksz96oW7z/0ODiS6c7506BrfI96x/s71HGsjnwIwJI0r67lxX7fkOZu/tcPhcdwrfvni2Df5VVuxV/OeAfwH+nkRJwP8BiyoVVDpVXQosBWhubq7Pv7WmInItOMbiysJ/X9f98yffM5l/vPCdQGLCuH55a8knXtWKwX6H28+fwZxJo1hwz+q8nYkG+TJLNwOuyzXnTuWe5/6AKw4xjbNkYf42zvkUyu/3pVePVfd4r9heQHuAS8v83TuBSWk/NyUfM6YoOQekrCv7B9du491TxvCeE8aw9vX2uh38AeIK555yTM6JL8WVnj1bIvE4l8+dzOVzJ/drsM1Xk1/odLViPtMGfu/0tg9gsareJSLfJ3cvoP4cCv8CcKKIHE9i4L8UuLwfn2caTPaA1BWJEc2R2//KT1qJo6jmKYepQQJkda8gFo+zpm0f86aNzXlGQcAVvn3xbIC8pZv9HWyzr9oB5t35rPXqqVO93QG8kvxn2Xv/q2pURK4Bfk6iDPR+VW2cpuumLNIHpF++upu7/6+tx2vC3ZNC/WQQXQceuvoMLrtvffddTTSeGFzX3Dg/q7VFjGvOPZHL507uHnQrmVpJv2rfsP3tfvfqsVbP3ultH8DjyX8+UIkvV9UngScr8dmmcYwZFuR7z7zWo6NnPQv6XJ7/w77EgStp/YpSg2tv+fNqpVb626vHWj17q7cU0OMUuGxS1QvKHpExeeS7Umzb3TGgBn9ItLD+4eo3ORTOXOpNH1xrIX/en1491urZe72lgL6d/OffAMdy9EjIy4DdlQrKmGyFrhTvX/OWt8EVSQSCPod4HL76VycxfJCPm366Ke/r0wf/VAfOWtwk1ddqHmv17L3eUkC/AhCR76hqc9pTj4uInQlsqqLQlSLAT16o/at/B/iXS+cw6S+GZgySa9/Yx2Mb/lTwvcV04PRaX+5GrNWz94rdCDZURKamfkhW7gytTEjGZNqxvwtXMit4UleKm3cdpB72dMWB6x/ewNb2QxkD5b9cdhoXzn5Hxmuza5V668BZr4ptG2Eqp9iNYP8P+KWIvEHiz+dxwN9VLCpj0mzaeaBHLjwci3OgK8LBrohHUZUuFNWcOe7vXXYqX5rfQev2t5kzaRRb/niwYfrf22YwbxW7EexnInIicEryoVeTB8UbU1HtnaFkT/pMkWicLy57iVC0tk70ckhUTeS7KcmX4542fnh3G4Rp44c31KBYC4vZjaqoFJCIDAFuAK5R1Q3AZBGxc4JNRbV3hnju1T34nJ4buOJARyiat9OlF/wOXPbuyQz25+/SU2yOu7deOsaUQ7EpoP8EXgTek/x5J/AwsKoSQZn6VM4NPamqHwc4XMQ5tV4TwHEcHm3dyeFI7ruSgCsDOp1j6k+xE8AJqvpxEbkMQFUPi0j97Ks3FVfODT3pVT/1QoFQNE6+vGjA51irY1Nziq0CCovIYJKpTRE5AfL+WTcNJl9b5mIPaMmWq+qnVhT7F8aX9cJLT2+ywd/UnGL/PN8G/AyYJCLLSLSDXlyxqExdKaVPfLp8J3k1jR5MOFZbi7spF2SVbKZkD/jZXUeXt+zo84RoTKX0mgISEQcYTWI38Bkk0p3XqWrPI4ZMQ+rLhp5CKaPVbfvQ2lnbzbDq5Z6btoYGXT539gnc+8s2/I5DKBZHVAmlLVDbDldTi3qdAJLHPy5W1eXAE1WIydSZUvvB9Laz98YVG2u2b3+udtOxuGb02x8acFlwz+qMXs62w9XUomIXgZ8Rka8CPwEOpR5U1T9XJCpTd0rZ0JOrB4zrCJt3HWDPwVDOsk+vOMCggEskFsd1pMfCdHZlT+qffW2QZkw1iRZxry0ib5L7QJipOV5eMc3NzdrSYi2I6l17Z6j7EJF0DhD0O3TVUPVP0Odw04dPYeaEEVx5/28zYg64wpPXnpV3cdf63JtaISIvZvVzA4pfBJ4O3AtsAFqB7wMz+hHMxSKyWUTiItIjKDOwjRkW5JYF03s8HgfPB//sw9ND0Th3/ez3XHn/b7mkuQl/2gviqmz548G8n2WbuUytK3YCeAD4SxIHw3+fxITQn0NiNpFYVH6+H59h6kSq2qdtd0d31c+k0YMZ7C/2j1/15NpYfDgS40gkzk9e2J7RqC11QpdV95h6VewawExVTb9ke05EejZoKZKqvgJge8kGvlS1j8YTVTGD/A6xuBKPa0108XQg42j1oCuoCD5HOJzVgM4VBxwySlStusfUs2IvwV4SkTNSP4jIXCpwTnAuIrJIRFpEpGXv3r3V+EpTJu2dIRY/soEjkXh3SeSRSJxIrDYGf58jZG1fQBzhyS+dyQ+uPJWgL/MCJaZxYllVQFbdY+pZsRPAacBvROQtEXkLWAucLiIvi8jGXG8QkWdEZFOOXxeWEqCqLlXVZlVtHjduXClvNR5btn4boWgNjPR5ROOKiBD0SUY/+mnjh3P2ScewZOHsjF71SxbOZslC619vBo5iU0DnlfrBqvqBUt9jBo72zhD3PtfmdRgZZk4YwWu7OzI6iLqOsPQTpzFycKBHtU6+0tZGatVsBrZizwPYWulATH1r7wyxeddBQJkxYSSbdx3AraF6foDf/6mDSFYK50gkzoSRg5k2fnj3YnX6wJ6rV731rzcDRbF3AGUlIh8jUU00DnhCRFpV9UNexGL6b2XrTr7yk9buvL7rCGht5PnT+X0OROMZk0DQFQ6FY2XtZmpMvfCkDk9Vf6qqTaoaVNXxNvjXr/bOUMbgD4nWCLU2+EMirlyLvkMDblm7mRpTLzy5AzADRy0fyu6QuOoPuEev6oEeLRoOhWM9WlNYeadpBDYBmH7yfvTPruVP+X8fPKm7QVuhRdz2zlDJ3UyLZe0gTC2rvUnGoWoAABNNSURBVK2Ypq7MmDAyoz1CtfldYfnfnUEwqyF/0Odw+dzJOdsxZD+W6mZa7vLOla07mXfns1x533rm3fksj7Xu7NfnGVNudgdgipZ+NQt0//5j75rI8pYdnsQUiylr3/gzt54/nTtWbelz981SupkWo1DLa7sTMLXCJgBTlJWtO1n8yEZcRwhFYskNVA7hWGJnr1fiwHeefo2gz+HWBdOZOXFknwfwcpZ35mp5besKptZYCsgU1N4Z4vnX9vLVhzcQisY5HI4R08Qu2kPhWFUH/+xjF9OFonHueGJLzeTa+3JKmjHVZhOAySuVw/67/2rx9CofErn+q88ufPxEMecQV0ul1hWMKSdLATWQUipS0nPYXvI7guMISxbOYt60sdy/+i1Cec6LrLUr7HKvKxhTbjYBNIhSd7rmymFXmyvww083M2PCyO7Bc8nCxFGLibWIOKrKkICvZo9dtLYRppbZBNAA+lKRkiuHXS2ugM8VliyczdknHZPx3AVzJjL9HSNo3f42cyaNYvTQgF1hG9NHNgE0gGIqUrLTQ2OGBfnKB07iW0+9WtV7gC+/fxrnnjI+74BuPXuMKR+bABpAoYqU9s4Qy9Zv497n2jJaJqz+w16Wv1jdjUuuwLmnjGf2pFE5n7faemPKyyaAASz9qv6ui2Zl9MC55aPTue/Xb/Afz79B6syW1OLqdQ+1etLgIaYwNODmfd5q640pL5sABqjExq0NuOIQ0zhLFs5mzY3z2bG/i007D/APj2/OW9rpVcFn0OdwKOsc3nTlqK233jzGHGX7AAag9s4Q1y9vJRRVDkdihKLKV5a3AolB9PbHt3he15+LCAUH8zHDglzS3JTx2CXNTUUP5Nabx5hMnkwAIrJERF4VkY0i8lMRyZ30NX2yeddBskvlo/HE47c8+jLhmLe1/dkG+52iNkq1d4Z69Bxa3rKjqL796esH1vPfmASv7gCeBmaq6izgNeAmj+IYoHJf3b/2p4M8uWl3lWPp3dVnTWXNjfN7reZJrQGkc0WK2v2b6721tHPYGC94dSLYL1Q1mvxxHdBU6PWmNLlaNLuO8M2nXvUoosIumD2hqDROrjWAQ+EYm3Yd6NN7a23nsDHVVgtrAJ8Bnsr3pIgsEpEWEWnZu3dvFcOqD6mDzNNTGfsPhbn4tCYCrsOQgEvQ56A1ekzjJ98zmWnjhxf12jHDgtzy0ek9Hr9j1ZZeUznWm8eYnipWBSQizwDH5njqZlVdmXzNzUAUWJbvc1R1KbAUoLm5uQaHMO+kNkW5AqGYsvhDJ7Pjz4d5cN227td8YPp4Orqi/Lqt3cNIe/I5wkNXz6X5+DElvW/mxJEMC7p0ho5WCxVbCmq9eYzJVLEJQFU/UOh5Efk0sAB4v6rawF6iXM3avvFkzxTPky97l/MP+hxE4G/fO4UH1m7lcFqJ52C/i9+Xv+Y/n6bRg4nGM/+4lJLKsd48xhzlyT4AETkPWAy8T1UPexFDvduxvwsPT2IsyAGW/90Z+H1u98D8n795K+M1fc2/p1I52Qe726BuTOm82gh2DxAEnhYRgHWq+jmPYqkb6ZuYhgZcjuRpi+wlnwPfvWROj9ROOQdtS+UYUx6eTACqOs2L761nR/P9wpFoDBTiNZI4G+x3ufG8k5k6bmhG6+Z05R60LZVjTP9ZK4galX0Aey0czpJPLB7n/CJKOW3QNqa22ARQg7JbHn/xnGm4UqMJf+Di0ybZwG5MHbIJoMbkanl8z3N/6FH5Ukv+dt4Ur0MwxvRBLWwEM2lytSzwuQ411r6nWykbuYwxtcXuAGpMrpYFoWgcF8jfKLm6Aq7w9Y/8JWdOG2uDvzF1zO4AqiRXy4Zc0lsWBJOF/tGYejb4D/I5XPSuiQR90t1C4dsXz+bT8463wd+YOif1tAm3ublZW1pavA6jZKWcY5uq/olEY3x86bqq9e95z9TRrH1jf4/HB/kd1tw4H8Dq7o2pUyLyoqo2Zz9uKaAKK+Uc22XrtnL7qi0EXCEUiVW1edsVc4/jpW0Huo+FTLnlo9O747SB35iBxVJAFVZsH/pl67Zy86ObCEfjdIZiVLPk3xUYMdhPwM2Mc2jQZebEkdULxBhTVXYHUGGF+tC3d4bYvOsgB7si3LZykyfxCXD3x+cwY8LIHnHG4mr98o0ZwGwCqLB8zctWt+3j+uWtPY5urCafwM++fHb3Yq41WTOmsdgEUAXZfXAA3vut//N08A/6hCULZ2dU8liTNWMai00AVZDe12fMsCDPv7aHRPLFmwqsIQGXH1x5KmefdEyP56xfjzGNwyaACssuAb3ktCZ+0rKdUNS78tu4KjMm2OKuMY3OJoAKylUCmn5cY7X4XcERCLhuztx+9h2KMaYxeHUi2B3AhUAc2AN8WlV3eRFLJaVKQFODv1eunX8il8+dnHOQL2WTmjFmYPFqH8ASVZ2lqnOAVcCtHsVRUblKQKst6BMunzuZMcOCzJ40qseVf+oOpSMU5UgkzuIVG3ttV2GMGRg8mQBU9WDaj0PxajW0nwr190mlVW5ZMJ2gr3q9/FPfFHSFQX6HJQtn503rFLtJzRgzMHm2BiAi/wR8EjgAnOtVHH2VL3XS3hli2fpt3PvcH/A5DuFYjHNOHMfTr+6tSlypmVRFeOKaMws2bCu0Sc0YM/BVbAIQkWeAY3M8dbOqrlTVm4GbReQm4BrgtjyfswhYBDB58uRKhVuSfP19Oo5Euf3xzYSTTXxCyR6e1Rr80wVdh0Phwj1E821Ss4VgYxqD591ARWQy8KSqzuzttbXSDfT51/bwuR+/xOG0AXZo0OVIuLoN3NI5knlIfNDn8JuvzS9qMLcqIGMGtnzdQD1ZAxCRE9N+vBB41Ys4+mJl606ufrAlY/AHiETVs8EfQLLODA5H46xp21fUe3MtEBtjBj6vqoC+JSKbRGQj8FfAdR7FUZJU6id7E1fAFeZMGlG1OHxOok9/uljWmcEK3PDIBqvoMcbk5VUV0EWqOjNZCnq+qu70Io5S7djfRTzH4ezhmPLbt96uSgwBV3jo6jOKeq0rVtFjjMnPzgMowbOv7O5e4K0GnyNkF5Be+u5JNB8/pvvYyOFBH0Gfg5uj0jSmVtFjjMnPJoAiLVu3le8921bV71x83skEfJn/i5a37KC9M8QFcyay5sb5/Piqufzma/O5++Nz8KfNAj6HgnsAjDHGegEVob0zxO2rtlTlu4b4HaJx5bbzZzBz4kgCrpNxTGNqo1aqa2dqgE+1ct686wAgzJgwwgZ/Y0xBNgEUYcf+LgKuEI5W5vODPgdQbj1/BjMnjOwux2zvDJW0UWvMsGDOFs/GGJOLpYCypLd3SP0+Eo1VNPfvCIAwPOjLKMdMbdRK5foH+R3bqGWMKRu7A0izbN1Wbl+1hYArhKJx4nHFEYjE6bEY218OdPcI7Yoc3U08b9rYjAHeTukyxlSKTQBJy9Zt5eZHEwezp6d6Uhf+5br+F+C+T57G/sMR/uHxzXSGjm4oS8/vp7NTuowxlWApIJKLvI9vrsp3fe59U3n/9GM595RjiGbtKbBGbMaYarIJgGRbZLfy/yn8rnDVWVMBy+8bY7xnKSASbZFjFWqK5wCDAi6xuLJk4SzL7xtjaoZNACSuxm/56HT+/tFNZT2Z5vPvm8pVZ00tOMBbft8Y4xWbAEgsAN+ysryDfyCZ7rEB3hhTqxp+Akiv/ikXvyt8+2Jrw2CMqW0NOwG07e5gdds+7lj1Slk+L+hz+M7Fsxkx2MeMCSNt8DfG1LyGnABuffRlHly3rSyfNSTgElflrotmsWD2hLJ8pjHGVEPDTQBtuzv6NPi7Qo8Tv4I+4QdXnmpX/MaYuuTpPgARuV5EVETGFvue9F49fdG6vW8Ht/h9Dl//8CkEXGFowGWQ32HJwtmcfdIxNvgbY+qSZ3cAIjKJxHGQRV+Ov304wrw7n8XvOETice66aBYXzJnY6/vSDz2fM2lUUd81p2kkr/ypg4Cb+V0XndZkdfvGmAHByxTQ3cBiYGWxb9jx9mGOjcQ5Qv7madmWrdvKbY9twhVBSZyolY8rcMuCv+TMaeOYNn54xsSR3qHTBn5jzEDgyQQgIhcCO1V1g0jhPpsisghYBOAfmdnrPl/ztJR//9XrfPOpVwGIJqv8H1yb/4bjHy+cyRVnHNf9sw32xpiBrGITgIg8Axyb46mbga+TSP/0SlWXAksBBk04MWMZtlDztGXrtnYP/r3xu8I/XDCDK+Ye1/uLjTFmgKjYBKCqH8j1uIi8EzgeSF39NwEvici7VfVPhT6zadQQfH4nYw0g1xV6Kd09A67w5LVnMW388KJeb4wxA0XVU0Cq+jLQncsRkbeAZlXd19t7Rw3x8/Mb5/e6CLtjf1dRsQR8Dt9eOMsGf2NMQ6q7fQDF5OWHBtxej3AMuMKTXzrTBn9jTMPyfAJQ1Snl/sxD4RiD/A5HIkcPVHck8Wuw39edPrLB3xjTyDyfACoh18JwwOew6pozORSOWQ2/McYwQE8Ey3fa1rTxw5k9aZQN/sYYQx3eAeTanJWLnbZljDGF1dUEUGorCNvIZYwx+dVVCmjH24c5EonTEYpyJBJn8YqNfW4KZ4wxja6uJgAhs21EqhWEMcaY0olqOU/CrSx38Aj1jRp/9AHVeHjvWy8Tj0X78HFjgV43n1WZxVQci6l4tRiXxVSccsZ0nKqOy36wriaAchKRFlVt9jqOdBZTcSym4tViXBZTcaoRU12lgIwxxpSPTQDGGNOgGnkCWOp1ADlYTMWxmIpXi3FZTMWpeEwNuwZgjDGNrpHvAIwxpqHZBGCMMQ2q4ScAEbleRFRExnodC4CI3CEiG0WkVUR+ISITaiCmJSLyajKun4rIqBqI6WIR2SwicRHxtHxPRM4Tkd+LSJuIfM3LWJLx3C8ie0Rkk9expIjIJBF5TkS2JP+/Xed1TAAiMkhEfisiG5Jx3e51TAAi4orI70RkVSW/p6EnABGZROJs4vwnxVffElWdpapzgFXArV4HBDwNzFTVWcBrwE0exwOwCfgb4HkvgxARF7gX+DAwHbhMRKZ7GRPwI+A8j2PIFgWuV9XpwBnAF2vgvxNACJivqrOBOcB5InKGxzEBXAe8UukvaegJALgbWAzUzEq4qh5M+3EoNRCbqv5CVVO7rdeROMfZU6r6iqr+3us4gHcDbar6hqqGgYeAC70MSFWfB/7sZQzZVPWPqvpS8vcdJAa3/J0cq0QTOpM/+pO/PP07JyJNwEeB+yr9XQ07AYjIhcBOVd3gdSzZROSfRGQ7cAW1cQeQ7jPAU14HUUMmAtvTft5BDQxstUxEpgDvAtZ7G0lCMt3SCuwBnlZVr+P6ZxIXpvHeXthfddUOulQi8gxwbI6nbga+TiL9U3WF4lLVlap6M3CziNwEXAPc5nVMydfcTOJWflml4yk2JlNfRGQYsAL4ctbdrmdUNQbMSa5t/VREZqqqJ+snIrIA2KOqL4rIOZX+vgE9AajqB3I9LiLvBI4HNogIJFIaL4nIu1X1T17FlcMy4EmqMAH0FpOIfBpYALxfq7R5pIT/Tl7aCUxK+7kp+ZjJIiJ+EoP/MlX9X6/jyaaqb4vIcyTWT7xaQJ8HXCAiHwEGASNE5MeqemUlvqwhU0Cq+rKqHqOqU5KH0u8ATq3G4N8bETkx7ccLgVe9iiVFRM4jcUt6gaoe9jqeGvMCcKKIHC8iAeBS4DGPY6o5krjS+iHwiqp+1+t4UkRkXKqqTUQGAx/Ew79zqnqTqjYlx6VLgWcrNfhDg04ANe5bIrJJRDaSSFHVQrncPcBw4OlkeeoPvA5IRD4mIjuA9wBPiMjPvYgjuTh+DfBzEguby1V1sxexpIjI/wBrgZNFZIeIfNbLeJLmAZ8A5if/DLUmr3K99g7gueTftxdIrAFUtPSyllgrCGOMaVB2B2CMMQ3KJgBjjGlQNgEYY0yDsgnAGGMalE0AxhjToGwCMCZJREaJyBeq8D1/XSON0EyDswnAmKNGAUVPAJLQl79Df02ic6gxnrJ9AMYkiUiqk+fvgeeAWcBoEh0i/15VVyYbmf2cRCOz04CPAJ8ErgT2kmgM96KqfltETiDRKnoccBi4GvgLEm2+DyR/XaSqr1fpX9GYDAO6F5AxJfoaiXMP5oiIDxiiqgeThwWtE5FUi4cTgU+p6joROR24CJhNYqJ4CXgx+bqlwOdU9Q8iMhf4V1Wdn/ycVar6SDX/5YzJZhOAMbkJ8A0ROZtEW96JwPjkc1tVdV3y9/OAlap6BDgiIo9Dd9fL9wIPJxsOAgSrFbwxxbAJwJjcriCRujlNVSMi8haJ7owAh4p4vwO8nTzZzZiaZIvAxhzVQaLpHcBIEn3ZIyJyLnBcnvesAc5Pni07jETL7NTJbm+KyMXQvWA8O8f3GOMZmwCMSVLVdmBN8jD1OUCziLxMYpE3Z4tgVX2BRPvnjSROSnuZxOIuJO4iPisiG4DNHD0q8iHghuSh3ydU6t/HmN5YFZAx/SQiw1S1U0SGkDikflHq/FtjapmtARjTf0uTG7sGAQ/Y4G/qhd0BGGNMg7I1AGOMaVA2ARhjTIOyCcAYYxqUTQDGGNOgbAIwxpgG9f8BVnznBk7ENAAAAAAASUVORK5CYII=\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%popd"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UDaSWepQu4mg",
        "outputId": "00465aad-6326-4b13-c881-f20c4331805e"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content\n",
            "popd -> /content\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!cd cgnn/examples; zip -r oqmd_fe_benchmark.zip oqmd_fe_benchmark"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GDZoP1XFuA2v",
        "outputId": "f0561a2b-74b2-4e65-ba03-48ce31f6eb46"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "  adding: oqmd_fe_benchmark/ (stored 0%)\n",
            "  adding: oqmd_fe_benchmark/run_oqmd.sh (deflated 41%)\n",
            "  adding: oqmd_fe_benchmark/model.pth (deflated 14%)\n",
            "  adding: oqmd_fe_benchmark/test_predictions.csv (deflated 59%)\n",
            "  adding: oqmd_fe_benchmark/history.csv (deflated 54%)\n",
            "  adding: oqmd_fe_benchmark/log.txt (deflated 86%)\n"
          ]
        }
      ]
    }
  ]
}