colab_sandbox.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.8.2-final"
    },
    "orig_nbformat": 2,
    "kernelspec": {
      "name": "python38264bitlanguagedetectionpipenv382f6b637c4e47f9bb779a88a8777172",
      "display_name": "Python 3.8.2 64-bit ('LanguageDetection': pipenv)"
    },
    "colab": {
      "name": "colab_sandbox.ipynb",
      "provenance": [],
      "include_colab_link": true
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/pablogadhi/LanguageDetection/blob/master/colab_sandbox.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Kn1jfXR_ugxS"
      },
      "source": [
        "#Environment Setup"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q0b-cXd1uSBm",
        "outputId": "af0fc838-4526-42fe-d822-2f35e1605d18",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 364
        }
      },
      "source": [
        "!nvidia-smi"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fri Oct 23 17:54:50 2020       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 455.23.05    Driver Version: 418.67       CUDA Version: 10.1     |\n",
            "|-------------------------------+----------------------+----------------------+\n",
            "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                               |                      |               MIG M. |\n",
            "|===============================+======================+======================|\n",
            "|   0  Tesla P100-PCIE...  Off  | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   33C    P0    26W / 250W |      0MiB / 16280MiB |      0%      Default |\n",
            "|                               |                      |                 ERR! |\n",
            "+-------------------------------+----------------------+----------------------+\n",
            "                                                                               \n",
            "+-----------------------------------------------------------------------------+\n",
            "| Processes:                                                                  |\n",
            "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
            "|        ID   ID                                                   Usage      |\n",
            "|=============================================================================|\n",
            "|  No running processes found                                                 |\n",
            "+-----------------------------------------------------------------------------+\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vdAOlqqQuNcY",
        "outputId": "ad444de3-642d-48a8-f978-21d0b57d7408",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 139
        }
      },
      "source": [
        "!git clone https://github.com/pablogadhi/LanguageDetection.git"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cloning into 'LanguageDetection'...\n",
            "remote: Enumerating objects: 205, done.\u001b[K\n",
            "remote: Counting objects: 100% (205/205), done.\u001b[K\n",
            "remote: Compressing objects: 100% (157/157), done.\u001b[K\n",
            "remote: Total 205 (delta 107), reused 119 (delta 43), pack-reused 0\u001b[K\n",
            "Receiving objects: 100% (205/205), 872.00 KiB | 6.46 MiB/s, done.\n",
            "Resolving deltas: 100% (107/107), done.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-x6l-fpMuNcj",
        "outputId": "b00ff95f-fc6f-467a-c63f-5166e50ced4e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "!pip install OpenNMT-py"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting OpenNMT-py\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/9f/20/40f8b722aa0e35e259c144b6ec2d684f1aea7de869cf586c67cfd6fe1c55/OpenNMT_py-1.2.0-py3-none-any.whl (195kB)\n",
            "\r\u001b[K     |█▊                              | 10kB 11.9MB/s eta 0:00:01\r\u001b[K     |███▍                            | 20kB 2.9MB/s eta 0:00:01\r\u001b[K     |█████                           | 30kB 3.6MB/s eta 0:00:01\r\u001b[K     |██████▊                         | 40kB 4.0MB/s eta 0:00:01\r\u001b[K     |████████▍                       | 51kB 3.4MB/s eta 0:00:01\r\u001b[K     |██████████                      | 61kB 3.9MB/s eta 0:00:01\r\u001b[K     |███████████▊                    | 71kB 4.1MB/s eta 0:00:01\r\u001b[K     |█████████████▍                  | 81kB 4.5MB/s eta 0:00:01\r\u001b[K     |███████████████                 | 92kB 4.8MB/s eta 0:00:01\r\u001b[K     |████████████████▉               | 102kB 4.6MB/s eta 0:00:01\r\u001b[K     |██████████████████▌             | 112kB 4.6MB/s eta 0:00:01\r\u001b[K     |████████████████████▏           | 122kB 4.6MB/s eta 0:00:01\r\u001b[K     |█████████████████████▉          | 133kB 4.6MB/s eta 0:00:01\r\u001b[K     |███████████████████████▌        | 143kB 4.6MB/s eta 0:00:01\r\u001b[K     |█████████████████████████▏      | 153kB 4.6MB/s eta 0:00:01\r\u001b[K     |██████████████████████████▉     | 163kB 4.6MB/s eta 0:00:01\r\u001b[K     |████████████████████████████▌   | 174kB 4.6MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████▏ | 184kB 4.6MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 194kB 4.6MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 204kB 4.6MB/s \n",
            "\u001b[?25hRequirement already satisfied: torch>=1.6.0 in /usr/local/lib/python3.6/dist-packages (from OpenNMT-py) (1.6.0+cu101)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.6/dist-packages (from OpenNMT-py) (4.41.1)\n",
            "Collecting configargparse\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/bb/79/3045743bb26ca2e44a1d317c37395462bfed82dbbd38e69a3280b63696ce/ConfigArgParse-1.2.3.tar.gz (42kB)\n",
            "\u001b[K     |████████████████████████████████| 51kB 5.9MB/s \n",
            "\u001b[?25hCollecting waitress\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/26/d1/5209fb8c764497a592363c47054436a515b47b8c3e4970ddd7184f088857/waitress-1.4.4-py2.py3-none-any.whl (58kB)\n",
            "\u001b[K     |████████████████████████████████| 61kB 5.6MB/s \n",
            "\u001b[?25hCollecting pyonmttok==1.*; platform_system == \"Linux\"\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/17/e5/d4544b969d863aed692f6d551010bd5988dfce4f53e1d7dd332702586c41/pyonmttok-1.21.0-cp36-cp36m-manylinux1_x86_64.whl (2.5MB)\n",
            "\u001b[K     |████████████████████████████████| 2.6MB 15.3MB/s \n",
            "\u001b[?25hRequirement already satisfied: six in /usr/local/lib/python3.6/dist-packages (from OpenNMT-py) (1.15.0)\n",
            "Requirement already satisfied: pyyaml in /usr/local/lib/python3.6/dist-packages (from OpenNMT-py) (3.13)\n",
            "Collecting torchtext==0.4.0\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/43/94/929d6bd236a4fb5c435982a7eb9730b78dcd8659acf328fd2ef9de85f483/torchtext-0.4.0-py3-none-any.whl (53kB)\n",
            "\u001b[K     |████████████████████████████████| 61kB 10.1MB/s \n",
            "\u001b[?25hRequirement already satisfied: tensorboard>=1.14 in /usr/local/lib/python3.6/dist-packages (from OpenNMT-py) (2.3.0)\n",
            "Requirement already satisfied: flask in /usr/local/lib/python3.6/dist-packages (from OpenNMT-py) (1.1.2)\n",
            "Requirement already satisfied: future in /usr/local/lib/python3.6/dist-packages (from OpenNMT-py) (0.16.0)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from torch>=1.6.0->OpenNMT-py) (1.18.5)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.6/dist-packages (from torchtext==0.4.0->OpenNMT-py) (2.23.0)\n",
            "Requirement already satisfied: tensorboard-plugin-wit>=1.6.0 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->OpenNMT-py) (1.7.0)\n",
            "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->OpenNMT-py) (3.2.2)\n",
            "Requirement already satisfied: setuptools>=41.0.0 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->OpenNMT-py) (50.3.0)\n",
            "Requirement already satisfied: wheel>=0.26; python_version >= \"3\" in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->OpenNMT-py) (0.35.1)\n",
            "Requirement already satisfied: google-auth-oauthlib<0.5,>=0.4.1 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->OpenNMT-py) (0.4.1)\n",
            "Requirement already satisfied: absl-py>=0.4 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->OpenNMT-py) (0.10.0)\n",
            "Requirement already satisfied: grpcio>=1.24.3 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->OpenNMT-py) (1.32.0)\n",
            "Requirement already satisfied: protobuf>=3.6.0 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->OpenNMT-py) (3.12.4)\n",
            "Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->OpenNMT-py) (1.0.1)\n",
            "Requirement already satisfied: google-auth<2,>=1.6.3 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->OpenNMT-py) (1.17.2)\n",
            "Requirement already satisfied: click>=5.1 in /usr/local/lib/python3.6/dist-packages (from flask->OpenNMT-py) (7.1.2)\n",
            "Requirement already satisfied: itsdangerous>=0.24 in /usr/local/lib/python3.6/dist-packages (from flask->OpenNMT-py) (1.1.0)\n",
            "Requirement already satisfied: Jinja2>=2.10.1 in /usr/local/lib/python3.6/dist-packages (from flask->OpenNMT-py) (2.11.2)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests->torchtext==0.4.0->OpenNMT-py) (3.0.4)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests->torchtext==0.4.0->OpenNMT-py) (2.10)\n",
            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from requests->torchtext==0.4.0->OpenNMT-py) (1.24.3)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests->torchtext==0.4.0->OpenNMT-py) (2020.6.20)\n",
            "Requirement already satisfied: importlib-metadata; python_version < \"3.8\" in /usr/local/lib/python3.6/dist-packages (from markdown>=2.6.8->tensorboard>=1.14->OpenNMT-py) (2.0.0)\n",
            "Requirement already satisfied: requests-oauthlib>=0.7.0 in /usr/local/lib/python3.6/dist-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard>=1.14->OpenNMT-py) (1.3.0)\n",
            "Requirement already satisfied: pyasn1-modules>=0.2.1 in /usr/local/lib/python3.6/dist-packages (from google-auth<2,>=1.6.3->tensorboard>=1.14->OpenNMT-py) (0.2.8)\n",
            "Requirement already satisfied: rsa<5,>=3.1.4; python_version >= \"3\" in /usr/local/lib/python3.6/dist-packages (from google-auth<2,>=1.6.3->tensorboard>=1.14->OpenNMT-py) (4.6)\n",
            "Requirement already satisfied: cachetools<5.0,>=2.0.0 in /usr/local/lib/python3.6/dist-packages (from google-auth<2,>=1.6.3->tensorboard>=1.14->OpenNMT-py) (4.1.1)\n",
            "Requirement already satisfied: MarkupSafe>=0.23 in /usr/local/lib/python3.6/dist-packages (from Jinja2>=2.10.1->flask->OpenNMT-py) (1.1.1)\n",
            "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.6/dist-packages (from importlib-metadata; python_version < \"3.8\"->markdown>=2.6.8->tensorboard>=1.14->OpenNMT-py) (3.2.0)\n",
            "Requirement already satisfied: oauthlib>=3.0.0 in /usr/local/lib/python3.6/dist-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<0.5,>=0.4.1->tensorboard>=1.14->OpenNMT-py) (3.1.0)\n",
            "Requirement already satisfied: pyasn1<0.5.0,>=0.4.6 in /usr/local/lib/python3.6/dist-packages (from pyasn1-modules>=0.2.1->google-auth<2,>=1.6.3->tensorboard>=1.14->OpenNMT-py) (0.4.8)\n",
            "Building wheels for collected packages: configargparse\n",
            "  Building wheel for configargparse (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for configargparse: filename=ConfigArgParse-1.2.3-cp36-none-any.whl size=19329 sha256=8b92a0d7e3026f9341231c188f7aee33a3cf8cc4045dc730ddd31dc8180bb1c2\n",
            "  Stored in directory: /root/.cache/pip/wheels/bd/d6/53/034032da9498bda2385cd50a51a289e88090b5da2d592b1fdf\n",
            "Successfully built configargparse\n",
            "Installing collected packages: configargparse, waitress, pyonmttok, torchtext, OpenNMT-py\n",
            "  Found existing installation: torchtext 0.3.1\n",
            "    Uninstalling torchtext-0.3.1:\n",
            "      Successfully uninstalled torchtext-0.3.1\n",
            "Successfully installed OpenNMT-py-1.2.0 configargparse-1.2.3 pyonmttok-1.21.0 torchtext-0.4.0 waitress-1.4.4\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_I-Sa3hWucdx",
        "outputId": "9b41a7c8-bf85-46a5-e72f-075f202ab214",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "%cd drive/My\\ Drive/Tesis"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n",
            "/content/drive/My Drive/Tesis\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2tQbz5Yhu1SY"
      },
      "source": [
        "#Training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CkOtW8aIvGzv",
        "outputId": "1515c7b0-0011-4635-e9b6-130f661c7624",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "!../../../LanguageDetection/onmt_train.sh data/multi-europarl models/multi-europarl-v3 57800 57801 4096 7225 models/multi-europarl-v3_step_28900.pt"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[2020-10-23 17:57:05,518 INFO] Loading checkpoint from models/multi-europarl-v3_step_28900.pt\n",
            "[2020-10-23 17:57:25,768 INFO] Loading vocab from checkpoint at models/multi-europarl-v3_step_28900.pt.\n",
            "[2020-10-23 17:57:26,293 INFO]  * src vocab size = 97861\n",
            "[2020-10-23 17:57:26,293 INFO]  * tgt vocab size = 97874\n",
            "[2020-10-23 17:57:26,293 INFO] Building model...\n",
            "[2020-10-23 17:57:38,312 INFO] NMTModel(\n",
            "  (encoder): TransformerEncoder(\n",
            "    (embeddings): Embeddings(\n",
            "      (make_embedding): Sequential(\n",
            "        (emb_luts): Elementwise(\n",
            "          (0): Embedding(97861, 512, padding_idx=1)\n",
            "        )\n",
            "        (pe): PositionalEncoding(\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "      )\n",
            "    )\n",
            "    (transformer): ModuleList(\n",
            "      (0): TransformerEncoderLayer(\n",
            "        (self_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (feed_forward): PositionwiseFeedForward(\n",
            "          (w_1): Linear(in_features=512, out_features=2048, bias=True)\n",
            "          (w_2): Linear(in_features=2048, out_features=512, bias=True)\n",
            "          (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "          (dropout_1): Dropout(p=0.1, inplace=False)\n",
            "          (relu): ReLU()\n",
            "          (dropout_2): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "        (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (dropout): Dropout(p=0.1, inplace=False)\n",
            "      )\n",
            "      (1): TransformerEncoderLayer(\n",
            "        (self_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (feed_forward): PositionwiseFeedForward(\n",
            "          (w_1): Linear(in_features=512, out_features=2048, bias=True)\n",
            "          (w_2): Linear(in_features=2048, out_features=512, bias=True)\n",
            "          (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "          (dropout_1): Dropout(p=0.1, inplace=False)\n",
            "          (relu): ReLU()\n",
            "          (dropout_2): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "        (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (dropout): Dropout(p=0.1, inplace=False)\n",
            "      )\n",
            "      (2): TransformerEncoderLayer(\n",
            "        (self_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (feed_forward): PositionwiseFeedForward(\n",
            "          (w_1): Linear(in_features=512, out_features=2048, bias=True)\n",
            "          (w_2): Linear(in_features=2048, out_features=512, bias=True)\n",
            "          (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "          (dropout_1): Dropout(p=0.1, inplace=False)\n",
            "          (relu): ReLU()\n",
            "          (dropout_2): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "        (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (dropout): Dropout(p=0.1, inplace=False)\n",
            "      )\n",
            "      (3): TransformerEncoderLayer(\n",
            "        (self_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (feed_forward): PositionwiseFeedForward(\n",
            "          (w_1): Linear(in_features=512, out_features=2048, bias=True)\n",
            "          (w_2): Linear(in_features=2048, out_features=512, bias=True)\n",
            "          (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "          (dropout_1): Dropout(p=0.1, inplace=False)\n",
            "          (relu): ReLU()\n",
            "          (dropout_2): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "        (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (dropout): Dropout(p=0.1, inplace=False)\n",
            "      )\n",
            "      (4): TransformerEncoderLayer(\n",
            "        (self_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (feed_forward): PositionwiseFeedForward(\n",
            "          (w_1): Linear(in_features=512, out_features=2048, bias=True)\n",
            "          (w_2): Linear(in_features=2048, out_features=512, bias=True)\n",
            "          (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "          (dropout_1): Dropout(p=0.1, inplace=False)\n",
            "          (relu): ReLU()\n",
            "          (dropout_2): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "        (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (dropout): Dropout(p=0.1, inplace=False)\n",
            "      )\n",
            "      (5): TransformerEncoderLayer(\n",
            "        (self_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (feed_forward): PositionwiseFeedForward(\n",
            "          (w_1): Linear(in_features=512, out_features=2048, bias=True)\n",
            "          (w_2): Linear(in_features=2048, out_features=512, bias=True)\n",
            "          (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "          (dropout_1): Dropout(p=0.1, inplace=False)\n",
            "          (relu): ReLU()\n",
            "          (dropout_2): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "        (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (dropout): Dropout(p=0.1, inplace=False)\n",
            "      )\n",
            "    )\n",
            "    (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "  )\n",
            "  (decoder): TransformerDecoder(\n",
            "    (embeddings): Embeddings(\n",
            "      (make_embedding): Sequential(\n",
            "        (emb_luts): Elementwise(\n",
            "          (0): Embedding(97874, 512, padding_idx=1)\n",
            "        )\n",
            "        (pe): PositionalEncoding(\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "      )\n",
            "    )\n",
            "    (transformer_layers): ModuleList(\n",
            "      (0): TransformerDecoderLayer(\n",
            "        (self_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (context_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (feed_forward): PositionwiseFeedForward(\n",
            "          (w_1): Linear(in_features=512, out_features=2048, bias=True)\n",
            "          (w_2): Linear(in_features=2048, out_features=512, bias=True)\n",
            "          (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "          (dropout_1): Dropout(p=0.1, inplace=False)\n",
            "          (relu): ReLU()\n",
            "          (dropout_2): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "        (layer_norm_1): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (layer_norm_2): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (drop): Dropout(p=0.1, inplace=False)\n",
            "      )\n",
            "      (1): TransformerDecoderLayer(\n",
            "        (self_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (context_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (feed_forward): PositionwiseFeedForward(\n",
            "          (w_1): Linear(in_features=512, out_features=2048, bias=True)\n",
            "          (w_2): Linear(in_features=2048, out_features=512, bias=True)\n",
            "          (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "          (dropout_1): Dropout(p=0.1, inplace=False)\n",
            "          (relu): ReLU()\n",
            "          (dropout_2): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "        (layer_norm_1): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (layer_norm_2): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (drop): Dropout(p=0.1, inplace=False)\n",
            "      )\n",
            "      (2): TransformerDecoderLayer(\n",
            "        (self_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (context_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (feed_forward): PositionwiseFeedForward(\n",
            "          (w_1): Linear(in_features=512, out_features=2048, bias=True)\n",
            "          (w_2): Linear(in_features=2048, out_features=512, bias=True)\n",
            "          (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "          (dropout_1): Dropout(p=0.1, inplace=False)\n",
            "          (relu): ReLU()\n",
            "          (dropout_2): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "        (layer_norm_1): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (layer_norm_2): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (drop): Dropout(p=0.1, inplace=False)\n",
            "      )\n",
            "      (3): TransformerDecoderLayer(\n",
            "        (self_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (context_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (feed_forward): PositionwiseFeedForward(\n",
            "          (w_1): Linear(in_features=512, out_features=2048, bias=True)\n",
            "          (w_2): Linear(in_features=2048, out_features=512, bias=True)\n",
            "          (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "          (dropout_1): Dropout(p=0.1, inplace=False)\n",
            "          (relu): ReLU()\n",
            "          (dropout_2): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "        (layer_norm_1): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (layer_norm_2): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (drop): Dropout(p=0.1, inplace=False)\n",
            "      )\n",
            "      (4): TransformerDecoderLayer(\n",
            "        (self_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (context_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (feed_forward): PositionwiseFeedForward(\n",
            "          (w_1): Linear(in_features=512, out_features=2048, bias=True)\n",
            "          (w_2): Linear(in_features=2048, out_features=512, bias=True)\n",
            "          (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "          (dropout_1): Dropout(p=0.1, inplace=False)\n",
            "          (relu): ReLU()\n",
            "          (dropout_2): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "        (layer_norm_1): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (layer_norm_2): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (drop): Dropout(p=0.1, inplace=False)\n",
            "      )\n",
            "      (5): TransformerDecoderLayer(\n",
            "        (self_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (context_attn): MultiHeadedAttention(\n",
            "          (linear_keys): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_values): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (linear_query): Linear(in_features=512, out_features=512, bias=True)\n",
            "          (softmax): Softmax(dim=-1)\n",
            "          (dropout): Dropout(p=0.1, inplace=False)\n",
            "          (final_linear): Linear(in_features=512, out_features=512, bias=True)\n",
            "        )\n",
            "        (feed_forward): PositionwiseFeedForward(\n",
            "          (w_1): Linear(in_features=512, out_features=2048, bias=True)\n",
            "          (w_2): Linear(in_features=2048, out_features=512, bias=True)\n",
            "          (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "          (dropout_1): Dropout(p=0.1, inplace=False)\n",
            "          (relu): ReLU()\n",
            "          (dropout_2): Dropout(p=0.1, inplace=False)\n",
            "        )\n",
            "        (layer_norm_1): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (layer_norm_2): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "        (drop): Dropout(p=0.1, inplace=False)\n",
            "      )\n",
            "    )\n",
            "    (layer_norm): LayerNorm((512,), eps=1e-06, elementwise_affine=True)\n",
            "  )\n",
            "  (generator): Sequential(\n",
            "    (0): Linear(in_features=512, out_features=97874, bias=True)\n",
            "    (1): Cast()\n",
            "    (2): LogSoftmax(dim=-1)\n",
            "  )\n",
            ")\n",
            "[2020-10-23 17:57:38,332 INFO] encoder: 69020160\n",
            "[2020-10-23 17:57:38,332 INFO] decoder: 125546066\n",
            "[2020-10-23 17:57:38,332 INFO] * number of parameters: 194566226\n",
            "[2020-10-23 17:57:39,660 INFO] Starting training on GPU: [0]\n",
            "[2020-10-23 17:57:39,660 INFO] Start training loop and validate every 57801 steps...\n",
            "[2020-10-23 17:57:39,660 INFO] Loading dataset from data/multi-europarl.train.0.pt\n",
            "[2020-10-23 17:58:01,399 INFO] number of examples: 427974\n",
            "[2020-10-23 17:58:59,272 INFO] Step 28950/57800; acc:  64.93; ppl:  4.60; xent: 1.53; lr: 0.00052; 4537/4522 tok/s;     80 sec\n",
            "[2020-10-23 17:59:53,044 INFO] Step 29000/57800; acc:  65.10; ppl:  4.55; xent: 1.51; lr: 0.00052; 6689/6521 tok/s;    133 sec\n",
            "[2020-10-23 18:00:46,358 INFO] Step 29050/57800; acc:  65.77; ppl:  4.37; xent: 1.47; lr: 0.00052; 6718/6664 tok/s;    187 sec\n",
            "[2020-10-23 18:01:41,313 INFO] Step 29100/57800; acc:  65.84; ppl:  4.36; xent: 1.47; lr: 0.00052; 6612/6670 tok/s;    242 sec\n",
            "[2020-10-23 18:02:35,825 INFO] Step 29150/57800; acc:  65.97; ppl:  4.32; xent: 1.46; lr: 0.00052; 6646/6690 tok/s;    296 sec\n",
            "[2020-10-23 18:03:29,979 INFO] Step 29200/57800; acc:  66.62; ppl:  4.15; xent: 1.42; lr: 0.00052; 6611/6688 tok/s;    350 sec\n",
            "[2020-10-23 18:04:24,300 INFO] Step 29250/57800; acc:  66.01; ppl:  4.26; xent: 1.45; lr: 0.00052; 6603/6555 tok/s;    405 sec\n",
            "[2020-10-23 18:05:18,226 INFO] Step 29300/57800; acc:  66.09; ppl:  4.24; xent: 1.44; lr: 0.00052; 6533/6640 tok/s;    459 sec\n",
            "[2020-10-23 18:06:12,156 INFO] Step 29350/57800; acc:  66.40; ppl:  4.16; xent: 1.43; lr: 0.00052; 6737/6658 tok/s;    512 sec\n",
            "[2020-10-23 18:07:06,440 INFO] Step 29400/57800; acc:  67.30; ppl:  3.99; xent: 1.38; lr: 0.00052; 6644/6677 tok/s;    567 sec\n",
            "[2020-10-23 18:08:00,381 INFO] Step 29450/57800; acc:  66.74; ppl:  4.08; xent: 1.41; lr: 0.00052; 6677/6656 tok/s;    621 sec\n",
            "[2020-10-23 18:08:55,165 INFO] Step 29500/57800; acc:  66.97; ppl:  4.02; xent: 1.39; lr: 0.00051; 6714/6621 tok/s;    676 sec\n",
            "[2020-10-23 18:09:49,485 INFO] Step 29550/57800; acc:  67.06; ppl:  4.01; xent: 1.39; lr: 0.00051; 6625/6699 tok/s;    730 sec\n",
            "[2020-10-23 18:10:42,776 INFO] Step 29600/57800; acc:  67.33; ppl:  3.92; xent: 1.37; lr: 0.00051; 6757/6711 tok/s;    783 sec\n",
            "[2020-10-23 18:11:36,433 INFO] Step 29650/57800; acc:  67.23; ppl:  3.94; xent: 1.37; lr: 0.00051; 6742/6667 tok/s;    837 sec\n",
            "[2020-10-23 18:12:30,591 INFO] Step 29700/57800; acc:  67.57; ppl:  3.87; xent: 1.35; lr: 0.00051; 6795/6689 tok/s;    891 sec\n",
            "[2020-10-23 18:13:23,984 INFO] Step 29750/57800; acc:  67.74; ppl:  3.83; xent: 1.34; lr: 0.00051; 6812/6745 tok/s;    944 sec\n",
            "[2020-10-23 18:14:18,267 INFO] Step 29800/57800; acc:  68.26; ppl:  3.74; xent: 1.32; lr: 0.00051; 6644/6756 tok/s;    999 sec\n",
            "[2020-10-23 18:15:13,122 INFO] Step 29850/57800; acc:  67.86; ppl:  3.81; xent: 1.34; lr: 0.00051; 6611/6618 tok/s;   1053 sec\n",
            "[2020-10-23 18:16:07,385 INFO] Step 29900/57800; acc:  68.20; ppl:  3.74; xent: 1.32; lr: 0.00051; 6760/6660 tok/s;   1108 sec\n",
            "[2020-10-23 18:17:01,836 INFO] Step 29950/57800; acc:  68.44; ppl:  3.70; xent: 1.31; lr: 0.00051; 6564/6644 tok/s;   1162 sec\n",
            "[2020-10-23 18:17:55,956 INFO] Step 30000/57800; acc:  68.32; ppl:  3.68; xent: 1.30; lr: 0.00051; 6652/6704 tok/s;   1216 sec\n",
            "[2020-10-23 18:18:50,858 INFO] Step 30050/57800; acc:  68.58; ppl:  3.65; xent: 1.30; lr: 0.00051; 6525/6734 tok/s;   1271 sec\n",
            "[2020-10-23 18:19:44,536 INFO] Step 30100/57800; acc:  68.92; ppl:  3.61; xent: 1.28; lr: 0.00051; 6761/6744 tok/s;   1325 sec\n",
            "[2020-10-23 18:20:38,458 INFO] Step 30150/57800; acc:  68.66; ppl:  3.62; xent: 1.29; lr: 0.00051; 6649/6782 tok/s;   1379 sec\n",
            "[2020-10-23 18:21:31,878 INFO] Step 30200/57800; acc:  69.00; ppl:  3.56; xent: 1.27; lr: 0.00051; 6767/6733 tok/s;   1432 sec\n",
            "[2020-10-23 18:22:25,861 INFO] Step 30250/57800; acc:  68.86; ppl:  3.59; xent: 1.28; lr: 0.00051; 6538/6687 tok/s;   1486 sec\n",
            "[2020-10-23 18:23:20,316 INFO] Step 30300/57800; acc:  69.41; ppl:  3.49; xent: 1.25; lr: 0.00051; 6687/6769 tok/s;   1541 sec\n",
            "[2020-10-23 18:24:13,960 INFO] Step 30350/57800; acc:  68.92; ppl:  3.56; xent: 1.27; lr: 0.00051; 6810/6651 tok/s;   1594 sec\n",
            "[2020-10-23 18:25:08,841 INFO] Step 30400/57800; acc:  69.11; ppl:  3.50; xent: 1.25; lr: 0.00051; 6528/6747 tok/s;   1649 sec\n",
            "[2020-10-23 18:25:19,917 INFO] Loading dataset from data/multi-europarl.train.1.pt\n",
            "[2020-10-23 18:25:40,292 INFO] number of examples: 426663\n",
            "[2020-10-23 18:26:28,106 INFO] Step 30450/57800; acc:  66.34; ppl:  4.21; xent: 1.44; lr: 0.00051; 4606/4646 tok/s;   1728 sec\n",
            "[2020-10-23 18:27:21,862 INFO] Step 30500/57800; acc:  65.67; ppl:  4.40; xent: 1.48; lr: 0.00051; 6760/6640 tok/s;   1782 sec\n",
            "[2020-10-23 18:28:15,689 INFO] Step 30550/57800; acc:  66.45; ppl:  4.24; xent: 1.44; lr: 0.00051; 6734/6777 tok/s;   1836 sec\n",
            "[2020-10-23 18:29:10,246 INFO] Step 30600/57800; acc:  65.67; ppl:  4.35; xent: 1.47; lr: 0.00051; 6697/6653 tok/s;   1891 sec\n",
            "[2020-10-23 18:30:04,890 INFO] Step 30650/57800; acc:  66.34; ppl:  4.20; xent: 1.43; lr: 0.00050; 6645/6746 tok/s;   1945 sec\n",
            "[2020-10-23 18:30:58,813 INFO] Step 30700/57800; acc:  66.91; ppl:  4.10; xent: 1.41; lr: 0.00050; 6785/6707 tok/s;   1999 sec\n",
            "[2020-10-23 18:31:52,579 INFO] Step 30750/57800; acc:  66.66; ppl:  4.14; xent: 1.42; lr: 0.00050; 6738/6675 tok/s;   2053 sec\n",
            "[2020-10-23 18:32:47,454 INFO] Step 30800/57800; acc:  66.98; ppl:  4.06; xent: 1.40; lr: 0.00050; 6513/6756 tok/s;   2108 sec\n",
            "[2020-10-23 18:33:42,104 INFO] Step 30850/57800; acc:  66.94; ppl:  4.06; xent: 1.40; lr: 0.00050; 6540/6662 tok/s;   2162 sec\n",
            "[2020-10-23 18:34:35,535 INFO] Step 30900/57800; acc:  66.97; ppl:  4.04; xent: 1.40; lr: 0.00050; 6682/6623 tok/s;   2216 sec\n",
            "[2020-10-23 18:35:29,612 INFO] Step 30950/57800; acc:  67.40; ppl:  3.94; xent: 1.37; lr: 0.00050; 6608/6697 tok/s;   2270 sec\n",
            "[2020-10-23 18:36:23,667 INFO] Step 31000/57800; acc:  67.13; ppl:  3.99; xent: 1.38; lr: 0.00050; 6716/6676 tok/s;   2324 sec\n",
            "[2020-10-23 18:37:17,823 INFO] Step 31050/57800; acc:  67.42; ppl:  3.92; xent: 1.37; lr: 0.00050; 6773/6665 tok/s;   2378 sec\n",
            "[2020-10-23 18:38:12,047 INFO] Step 31100/57800; acc:  67.52; ppl:  3.90; xent: 1.36; lr: 0.00050; 6623/6685 tok/s;   2432 sec\n",
            "[2020-10-23 18:39:06,932 INFO] Step 31150/57800; acc:  68.01; ppl:  3.78; xent: 1.33; lr: 0.00050; 6564/6756 tok/s;   2487 sec\n",
            "[2020-10-23 18:40:01,546 INFO] Step 31200/57800; acc:  68.23; ppl:  3.74; xent: 1.32; lr: 0.00050; 6601/6775 tok/s;   2542 sec\n",
            "[2020-10-23 18:40:55,021 INFO] Step 31250/57800; acc:  67.72; ppl:  3.84; xent: 1.35; lr: 0.00050; 6786/6586 tok/s;   2595 sec\n",
            "[2020-10-23 18:41:49,669 INFO] Step 31300/57800; acc:  67.92; ppl:  3.81; xent: 1.34; lr: 0.00050; 6603/6667 tok/s;   2650 sec\n",
            "[2020-10-23 18:42:43,756 INFO] Step 31350/57800; acc:  68.14; ppl:  3.76; xent: 1.32; lr: 0.00050; 6725/6704 tok/s;   2704 sec\n",
            "[2020-10-23 18:43:38,615 INFO] Step 31400/57800; acc:  68.23; ppl:  3.71; xent: 1.31; lr: 0.00050; 6579/6670 tok/s;   2759 sec\n",
            "[2020-10-23 18:44:33,434 INFO] Step 31450/57800; acc:  68.79; ppl:  3.62; xent: 1.29; lr: 0.00050; 6480/6630 tok/s;   2814 sec\n",
            "[2020-10-23 18:45:26,691 INFO] Step 31500/57800; acc:  68.26; ppl:  3.72; xent: 1.31; lr: 0.00050; 6761/6680 tok/s;   2867 sec\n",
            "[2020-10-23 18:46:20,560 INFO] Step 31550/57800; acc:  68.48; ppl:  3.67; xent: 1.30; lr: 0.00050; 6757/6674 tok/s;   2921 sec\n",
            "[2020-10-23 18:47:14,498 INFO] Step 31600/57800; acc:  68.63; ppl:  3.64; xent: 1.29; lr: 0.00050; 6743/6667 tok/s;   2975 sec\n",
            "[2020-10-23 18:48:08,168 INFO] Step 31650/57800; acc:  68.64; ppl:  3.65; xent: 1.29; lr: 0.00050; 6755/6648 tok/s;   3029 sec\n",
            "[2020-10-23 18:49:01,188 INFO] Step 31700/57800; acc:  68.56; ppl:  3.63; xent: 1.29; lr: 0.00050; 6809/6676 tok/s;   3082 sec\n",
            "[2020-10-23 18:49:56,287 INFO] Step 31750/57800; acc:  68.46; ppl:  3.63; xent: 1.29; lr: 0.00050; 6613/6532 tok/s;   3137 sec\n",
            "[2020-10-23 18:50:49,425 INFO] Step 31800/57800; acc:  69.22; ppl:  3.51; xent: 1.26; lr: 0.00050; 6734/6737 tok/s;   3190 sec\n",
            "[2020-10-23 18:51:43,495 INFO] Step 31850/57800; acc:  69.34; ppl:  3.49; xent: 1.25; lr: 0.00050; 6721/6712 tok/s;   3244 sec\n",
            "[2020-10-23 18:52:37,050 INFO] Step 31900/57800; acc:  69.70; ppl:  3.42; xent: 1.23; lr: 0.00049; 6691/6744 tok/s;   3297 sec\n",
            "[2020-10-23 18:52:59,821 INFO] Loading dataset from data/multi-europarl.train.2.pt\n",
            "[2020-10-23 18:53:20,621 INFO] number of examples: 426350\n",
            "[2020-10-23 18:53:57,018 INFO] Step 31950/57800; acc:  67.18; ppl:  3.99; xent: 1.38; lr: 0.00049; 4420/4591 tok/s;   3377 sec\n",
            "[2020-10-23 18:54:51,367 INFO] Step 32000/57800; acc:  65.52; ppl:  4.40; xent: 1.48; lr: 0.00049; 6666/6694 tok/s;   3432 sec\n",
            "[2020-10-23 18:55:46,041 INFO] Step 32050/57800; acc:  66.23; ppl:  4.25; xent: 1.45; lr: 0.00049; 6693/6712 tok/s;   3486 sec\n",
            "[2020-10-23 18:56:40,299 INFO] Step 32100/57800; acc:  66.57; ppl:  4.16; xent: 1.43; lr: 0.00049; 6641/6663 tok/s;   3541 sec\n",
            "[2020-10-23 18:57:34,157 INFO] Step 32150/57800; acc:  66.74; ppl:  4.13; xent: 1.42; lr: 0.00049; 6686/6699 tok/s;   3594 sec\n",
            "[2020-10-23 18:58:28,316 INFO] Step 32200/57800; acc:  66.57; ppl:  4.15; xent: 1.42; lr: 0.00049; 6757/6654 tok/s;   3649 sec\n",
            "[2020-10-23 18:59:22,422 INFO] Step 32250/57800; acc:  66.80; ppl:  4.08; xent: 1.41; lr: 0.00049; 6697/6603 tok/s;   3703 sec\n",
            "[2020-10-23 19:00:16,600 INFO] Step 32300/57800; acc:  67.02; ppl:  4.03; xent: 1.39; lr: 0.00049; 6747/6757 tok/s;   3757 sec\n",
            "[2020-10-23 19:01:10,328 INFO] Step 32350/57800; acc:  66.72; ppl:  4.07; xent: 1.40; lr: 0.00049; 6747/6699 tok/s;   3811 sec\n",
            "[2020-10-23 19:02:04,922 INFO] Step 32400/57800; acc:  67.20; ppl:  3.98; xent: 1.38; lr: 0.00049; 6566/6668 tok/s;   3865 sec\n",
            "[2020-10-23 19:02:58,759 INFO] Step 32450/57800; acc:  66.79; ppl:  4.04; xent: 1.40; lr: 0.00049; 6676/6581 tok/s;   3919 sec\n",
            "[2020-10-23 19:03:53,072 INFO] Step 32500/57800; acc:  67.53; ppl:  3.90; xent: 1.36; lr: 0.00049; 6583/6641 tok/s;   3973 sec\n",
            "[2020-10-23 19:04:47,558 INFO] Step 32550/57800; acc:  67.43; ppl:  3.92; xent: 1.37; lr: 0.00049; 6726/6656 tok/s;   4028 sec\n",
            "[2020-10-23 19:05:42,408 INFO] Step 32600/57800; acc:  67.87; ppl:  3.82; xent: 1.34; lr: 0.00049; 6638/6697 tok/s;   4083 sec\n",
            "[2020-10-23 19:06:35,992 INFO] Step 32650/57800; acc:  67.77; ppl:  3.83; xent: 1.34; lr: 0.00049; 6763/6637 tok/s;   4136 sec\n",
            "[2020-10-23 19:07:29,425 INFO] Step 32700/57800; acc:  67.95; ppl:  3.79; xent: 1.33; lr: 0.00049; 6836/6714 tok/s;   4190 sec\n",
            "[2020-10-23 19:08:23,837 INFO] Step 32750/57800; acc:  68.10; ppl:  3.76; xent: 1.33; lr: 0.00049; 6496/6574 tok/s;   4244 sec\n",
            "[2020-10-23 19:09:17,331 INFO] Step 32800/57800; acc:  68.78; ppl:  3.65; xent: 1.29; lr: 0.00049; 6648/6716 tok/s;   4298 sec\n",
            "[2020-10-23 19:10:11,503 INFO] Step 32850/57800; acc:  68.50; ppl:  3.68; xent: 1.30; lr: 0.00049; 6566/6584 tok/s;   4352 sec\n",
            "[2020-10-23 19:11:05,716 INFO] Step 32900/57800; acc:  68.40; ppl:  3.68; xent: 1.30; lr: 0.00049; 6571/6725 tok/s;   4406 sec\n",
            "[2020-10-23 19:12:00,108 INFO] Step 32950/57800; acc:  68.38; ppl:  3.68; xent: 1.30; lr: 0.00049; 6725/6689 tok/s;   4460 sec\n",
            "[2020-10-23 19:12:54,679 INFO] Step 33000/57800; acc:  68.25; ppl:  3.72; xent: 1.31; lr: 0.00049; 6638/6641 tok/s;   4515 sec\n",
            "[2020-10-23 19:13:48,327 INFO] Step 33050/57800; acc:  69.23; ppl:  3.54; xent: 1.26; lr: 0.00049; 6718/6794 tok/s;   4569 sec\n",
            "[2020-10-23 19:14:42,089 INFO] Step 33100/57800; acc:  68.93; ppl:  3.57; xent: 1.27; lr: 0.00049; 6726/6792 tok/s;   4622 sec\n",
            "[2020-10-23 19:15:36,446 INFO] Step 33150/57800; acc:  69.07; ppl:  3.55; xent: 1.27; lr: 0.00049; 6803/6699 tok/s;   4677 sec\n",
            "[2020-10-23 19:16:29,791 INFO] Step 33200/57800; acc:  69.18; ppl:  3.52; xent: 1.26; lr: 0.00049; 6759/6714 tok/s;   4730 sec\n",
            "[2020-10-23 19:17:23,639 INFO] Step 33250/57800; acc:  69.37; ppl:  3.49; xent: 1.25; lr: 0.00048; 6742/6715 tok/s;   4784 sec\n",
            "[2020-10-23 19:18:18,281 INFO] Step 33300/57800; acc:  69.47; ppl:  3.48; xent: 1.25; lr: 0.00048; 6582/6658 tok/s;   4839 sec\n",
            "[2020-10-23 19:19:12,490 INFO] Step 33350/57800; acc:  69.46; ppl:  3.47; xent: 1.25; lr: 0.00048; 6619/6643 tok/s;   4893 sec\n",
            "[2020-10-23 19:20:06,983 INFO] Step 33400/57800; acc:  69.04; ppl:  3.52; xent: 1.26; lr: 0.00048; 6531/6590 tok/s;   4947 sec\n",
            "[2020-10-23 19:20:39,237 INFO] Loading dataset from data/multi-europarl.train.3.pt\n",
            "[2020-10-23 19:21:02,503 INFO] number of examples: 426711\n",
            "[2020-10-23 19:21:28,173 INFO] Step 33450/57800; acc:  68.17; ppl:  3.75; xent: 1.32; lr: 0.00048; 4467/4436 tok/s;   5029 sec\n",
            "[2020-10-23 19:22:22,777 INFO] Step 33500/57800; acc:  66.10; ppl:  4.29; xent: 1.46; lr: 0.00048; 6551/6671 tok/s;   5083 sec\n",
            "[2020-10-23 19:23:16,510 INFO] Step 33550/57800; acc:  66.86; ppl:  4.12; xent: 1.42; lr: 0.00048; 6767/6756 tok/s;   5137 sec\n",
            "[2020-10-23 19:24:10,286 INFO] Step 33600/57800; acc:  66.28; ppl:  4.19; xent: 1.43; lr: 0.00048; 6832/6674 tok/s;   5191 sec\n",
            "[2020-10-23 19:25:04,740 INFO] Step 33650/57800; acc:  66.58; ppl:  4.12; xent: 1.42; lr: 0.00048; 6648/6613 tok/s;   5245 sec\n",
            "[2020-10-23 19:25:58,570 INFO] Step 33700/57800; acc:  67.13; ppl:  4.02; xent: 1.39; lr: 0.00048; 6799/6709 tok/s;   5299 sec\n",
            "[2020-10-23 19:26:53,344 INFO] Step 33750/57800; acc:  67.09; ppl:  4.02; xent: 1.39; lr: 0.00048; 6508/6623 tok/s;   5354 sec\n",
            "[2020-10-23 19:27:47,549 INFO] Step 33800/57800; acc:  66.84; ppl:  4.06; xent: 1.40; lr: 0.00048; 6552/6641 tok/s;   5408 sec\n",
            "[2020-10-23 19:28:42,119 INFO] Step 33850/57800; acc:  67.22; ppl:  3.97; xent: 1.38; lr: 0.00048; 6556/6726 tok/s;   5462 sec\n",
            "[2020-10-23 19:29:36,092 INFO] Step 33900/57800; acc:  68.11; ppl:  3.80; xent: 1.33; lr: 0.00048; 6644/6698 tok/s;   5516 sec\n",
            "[2020-10-23 19:30:29,594 INFO] Step 33950/57800; acc:  67.36; ppl:  3.95; xent: 1.37; lr: 0.00048; 6724/6530 tok/s;   5570 sec\n",
            "[2020-10-23 19:31:23,774 INFO] Step 34000/57800; acc:  67.96; ppl:  3.81; xent: 1.34; lr: 0.00048; 6636/6511 tok/s;   5624 sec\n",
            "[2020-10-23 19:32:18,998 INFO] Step 34050/57800; acc:  68.12; ppl:  3.78; xent: 1.33; lr: 0.00048; 6550/6713 tok/s;   5679 sec\n",
            "[2020-10-23 19:33:13,233 INFO] Step 34100/57800; acc:  67.86; ppl:  3.81; xent: 1.34; lr: 0.00048; 6733/6711 tok/s;   5734 sec\n",
            "[2020-10-23 19:34:07,796 INFO] Step 34150/57800; acc:  68.41; ppl:  3.70; xent: 1.31; lr: 0.00048; 6540/6775 tok/s;   5788 sec\n",
            "[2020-10-23 19:35:01,530 INFO] Step 34200/57800; acc:  68.89; ppl:  3.62; xent: 1.29; lr: 0.00048; 6711/6667 tok/s;   5842 sec\n",
            "[2020-10-23 19:35:56,754 INFO] Step 34250/57800; acc:  68.27; ppl:  3.71; xent: 1.31; lr: 0.00048; 6504/6722 tok/s;   5897 sec\n",
            "[2020-10-23 19:36:52,174 INFO] Step 34300/57800; acc:  68.63; ppl:  3.64; xent: 1.29; lr: 0.00048; 6527/6731 tok/s;   5953 sec\n",
            "[2020-10-23 19:37:45,669 INFO] Step 34350/57800; acc:  68.66; ppl:  3.65; xent: 1.29; lr: 0.00048; 6782/6685 tok/s;   6006 sec\n",
            "[2020-10-23 19:38:40,620 INFO] Step 34400/57800; acc:  68.50; ppl:  3.66; xent: 1.30; lr: 0.00048; 6675/6671 tok/s;   6061 sec\n",
            "[2020-10-23 19:39:34,334 INFO] Step 34450/57800; acc:  69.40; ppl:  3.51; xent: 1.26; lr: 0.00048; 6735/6660 tok/s;   6115 sec\n",
            "[2020-10-23 19:40:28,644 INFO] Step 34500/57800; acc:  69.02; ppl:  3.55; xent: 1.27; lr: 0.00048; 6681/6606 tok/s;   6169 sec\n",
            "[2020-10-23 19:41:22,813 INFO] Step 34550/57800; acc:  68.47; ppl:  3.63; xent: 1.29; lr: 0.00048; 6541/6582 tok/s;   6223 sec\n",
            "[2020-10-23 19:42:16,327 INFO] Step 34600/57800; acc:  69.41; ppl:  3.47; xent: 1.24; lr: 0.00048; 6835/6746 tok/s;   6277 sec\n",
            "[2020-10-23 19:43:10,542 INFO] Step 34650/57800; acc:  69.00; ppl:  3.53; xent: 1.26; lr: 0.00047; 6543/6756 tok/s;   6331 sec\n",
            "[2020-10-23 19:44:04,486 INFO] Step 34700/57800; acc:  68.88; ppl:  3.55; xent: 1.27; lr: 0.00047; 6791/6635 tok/s;   6385 sec\n",
            "[2020-10-23 19:44:58,815 INFO] Step 34750/57800; acc:  69.23; ppl:  3.50; xent: 1.25; lr: 0.00047; 6710/6622 tok/s;   6439 sec\n",
            "[2020-10-23 19:45:52,510 INFO] Step 34800/57800; acc:  69.35; ppl:  3.46; xent: 1.24; lr: 0.00047; 6679/6736 tok/s;   6493 sec\n",
            "[2020-10-23 19:46:46,374 INFO] Step 34850/57800; acc:  69.33; ppl:  3.46; xent: 1.24; lr: 0.00047; 6672/6698 tok/s;   6547 sec\n",
            "[2020-10-23 19:47:40,250 INFO] Step 34900/57800; acc:  70.00; ppl:  3.37; xent: 1.21; lr: 0.00047; 6699/6707 tok/s;   6601 sec\n",
            "[2020-10-23 19:48:18,045 INFO] Loading dataset from data/multi-europarl.train.4.pt\n",
            "[2020-10-23 19:48:36,919 INFO] number of examples: 424031\n",
            "[2020-10-23 19:48:57,808 INFO] Step 34950/57800; acc:  68.96; ppl:  3.59; xent: 1.28; lr: 0.00047; 4674/4652 tok/s;   6678 sec\n",
            "[2020-10-23 19:49:52,317 INFO] Step 35000/57800; acc:  66.27; ppl:  4.22; xent: 1.44; lr: 0.00047; 6625/6707 tok/s;   6733 sec\n",
            "[2020-10-23 19:50:45,912 INFO] Step 35050/57800; acc:  66.51; ppl:  4.17; xent: 1.43; lr: 0.00047; 6717/6543 tok/s;   6786 sec\n",
            "[2020-10-23 19:51:40,303 INFO] Step 35100/57800; acc:  67.07; ppl:  4.03; xent: 1.39; lr: 0.00047; 6682/6715 tok/s;   6841 sec\n",
            "[2020-10-23 19:52:33,364 INFO] Step 35150/57800; acc:  67.27; ppl:  3.99; xent: 1.38; lr: 0.00047; 6910/6714 tok/s;   6894 sec\n",
            "[2020-10-23 19:53:28,380 INFO] Step 35200/57800; acc:  66.90; ppl:  4.05; xent: 1.40; lr: 0.00047; 6608/6710 tok/s;   6949 sec\n",
            "[2020-10-23 19:54:23,112 INFO] Step 35250/57800; acc:  67.18; ppl:  3.97; xent: 1.38; lr: 0.00047; 6584/6603 tok/s;   7003 sec\n",
            "[2020-10-23 19:55:17,773 INFO] Step 35300/57800; acc:  67.43; ppl:  3.90; xent: 1.36; lr: 0.00047; 6532/6741 tok/s;   7058 sec\n",
            "[2020-10-23 19:56:12,285 INFO] Step 35350/57800; acc:  67.61; ppl:  3.89; xent: 1.36; lr: 0.00047; 6556/6661 tok/s;   7113 sec\n",
            "[2020-10-23 19:57:06,119 INFO] Step 35400/57800; acc:  67.47; ppl:  3.92; xent: 1.37; lr: 0.00047; 6834/6613 tok/s;   7166 sec\n",
            "[2020-10-23 19:58:00,535 INFO] Step 35450/57800; acc:  68.29; ppl:  3.76; xent: 1.33; lr: 0.00047; 6717/6727 tok/s;   7221 sec\n",
            "[2020-10-23 19:58:54,555 INFO] Step 35500/57800; acc:  67.81; ppl:  3.82; xent: 1.34; lr: 0.00047; 6688/6613 tok/s;   7275 sec\n",
            "[2020-10-23 19:59:48,707 INFO] Step 35550/57800; acc:  68.32; ppl:  3.73; xent: 1.32; lr: 0.00047; 6735/6613 tok/s;   7329 sec\n",
            "[2020-10-23 20:00:42,766 INFO] Step 35600/57800; acc:  68.20; ppl:  3.74; xent: 1.32; lr: 0.00047; 6693/6688 tok/s;   7383 sec\n",
            "[2020-10-23 20:01:37,110 INFO] Step 35650/57800; acc:  68.33; ppl:  3.72; xent: 1.31; lr: 0.00047; 6590/6674 tok/s;   7437 sec\n",
            "[2020-10-23 20:02:31,410 INFO] Step 35700/57800; acc:  68.87; ppl:  3.61; xent: 1.28; lr: 0.00047; 6645/6794 tok/s;   7492 sec\n",
            "[2020-10-23 20:03:25,892 INFO] Step 35750/57800; acc:  68.70; ppl:  3.62; xent: 1.29; lr: 0.00047; 6595/6743 tok/s;   7546 sec\n",
            "[2020-10-23 20:04:20,002 INFO] Step 35800/57800; acc:  68.89; ppl:  3.60; xent: 1.28; lr: 0.00047; 6733/6664 tok/s;   7600 sec\n",
            "[2020-10-23 20:05:13,689 INFO] Step 35850/57800; acc:  68.66; ppl:  3.63; xent: 1.29; lr: 0.00047; 6757/6718 tok/s;   7654 sec\n",
            "[2020-10-23 20:06:07,505 INFO] Step 35900/57800; acc:  69.12; ppl:  3.55; xent: 1.27; lr: 0.00047; 6720/6675 tok/s;   7708 sec\n",
            "[2020-10-23 20:07:01,175 INFO] Step 35950/57800; acc:  68.72; ppl:  3.62; xent: 1.29; lr: 0.00047; 6738/6707 tok/s;   7762 sec\n",
            "[2020-10-23 20:07:55,757 INFO] Step 36000/57800; acc:  69.37; ppl:  3.49; xent: 1.25; lr: 0.00047; 6650/6651 tok/s;   7816 sec\n",
            "[2020-10-23 20:08:49,419 INFO] Step 36050/57800; acc:  69.73; ppl:  3.43; xent: 1.23; lr: 0.00047; 6720/6658 tok/s;   7870 sec\n",
            "[2020-10-23 20:09:44,415 INFO] Step 36100/57800; acc:  69.02; ppl:  3.53; xent: 1.26; lr: 0.00047; 6411/6567 tok/s;   7925 sec\n",
            "[2020-10-23 20:10:12,113 INFO] Saving checkpoint models/multi-europarl-v3_step_36125.pt\n",
            "[2020-10-23 20:11:12,619 INFO] Step 36150/57800; acc:  70.18; ppl:  3.38; xent: 1.22; lr: 0.00046; 4105/4091 tok/s;   8013 sec\n",
            "[2020-10-23 20:12:07,819 INFO] Step 36200/57800; acc:  69.94; ppl:  3.40; xent: 1.22; lr: 0.00046; 6560/6665 tok/s;   8068 sec\n",
            "[2020-10-23 20:13:01,536 INFO] Step 36250/57800; acc:  69.36; ppl:  3.48; xent: 1.25; lr: 0.00046; 6581/6518 tok/s;   8122 sec\n",
            "[2020-10-23 20:13:56,197 INFO] Step 36300/57800; acc:  69.40; ppl:  3.46; xent: 1.24; lr: 0.00046; 6587/6710 tok/s;   8177 sec\n",
            "[2020-10-23 20:14:50,056 INFO] Step 36350/57800; acc:  69.81; ppl:  3.39; xent: 1.22; lr: 0.00046; 6759/6607 tok/s;   8230 sec\n",
            "[2020-10-23 20:15:44,298 INFO] Step 36400/57800; acc:  70.01; ppl:  3.36; xent: 1.21; lr: 0.00046; 6714/6690 tok/s;   8285 sec\n",
            "[2020-10-23 20:16:18,537 INFO] Loading dataset from data/multi-europarl.train.5.pt\n",
            "[2020-10-23 20:16:42,178 INFO] number of examples: 426172\n",
            "[2020-10-23 20:17:08,492 INFO] Step 36450/57800; acc:  68.04; ppl:  3.74; xent: 1.32; lr: 0.00046; 4201/4449 tok/s;   8369 sec\n",
            "[2020-10-23 20:18:02,597 INFO] Step 36500/57800; acc:  66.78; ppl:  4.09; xent: 1.41; lr: 0.00046; 6624/6671 tok/s;   8423 sec\n",
            "[2020-10-23 20:18:57,319 INFO] Step 36550/57800; acc:  66.67; ppl:  4.09; xent: 1.41; lr: 0.00046; 6585/6756 tok/s;   8478 sec\n",
            "[2020-10-23 20:19:51,678 INFO] Step 36600/57800; acc:  66.93; ppl:  4.03; xent: 1.39; lr: 0.00046; 6677/6690 tok/s;   8532 sec\n",
            "[2020-10-23 20:20:46,030 INFO] Step 36650/57800; acc:  67.69; ppl:  3.90; xent: 1.36; lr: 0.00046; 6665/6710 tok/s;   8586 sec\n",
            "[2020-10-23 20:21:39,713 INFO] Step 36700/57800; acc:  67.61; ppl:  3.90; xent: 1.36; lr: 0.00046; 6668/6659 tok/s;   8640 sec\n",
            "[2020-10-23 20:22:33,858 INFO] Step 36750/57800; acc:  67.97; ppl:  3.81; xent: 1.34; lr: 0.00046; 6648/6773 tok/s;   8694 sec\n",
            "[2020-10-23 20:23:28,001 INFO] Step 36800/57800; acc:  67.78; ppl:  3.85; xent: 1.35; lr: 0.00046; 6814/6670 tok/s;   8748 sec\n",
            "[2020-10-23 20:24:22,205 INFO] Step 36850/57800; acc:  68.02; ppl:  3.79; xent: 1.33; lr: 0.00046; 6698/6706 tok/s;   8803 sec\n",
            "[2020-10-23 20:25:15,510 INFO] Step 36900/57800; acc:  68.29; ppl:  3.74; xent: 1.32; lr: 0.00046; 6926/6627 tok/s;   8856 sec\n",
            "[2020-10-23 20:26:09,951 INFO] Step 36950/57800; acc:  68.26; ppl:  3.73; xent: 1.32; lr: 0.00046; 6628/6669 tok/s;   8910 sec\n",
            "[2020-10-23 20:27:03,990 INFO] Step 37000/57800; acc:  68.22; ppl:  3.74; xent: 1.32; lr: 0.00046; 6747/6627 tok/s;   8964 sec\n",
            "[2020-10-23 20:27:58,961 INFO] Step 37050/57800; acc:  68.18; ppl:  3.72; xent: 1.31; lr: 0.00046; 6571/6718 tok/s;   9019 sec\n",
            "[2020-10-23 20:28:52,590 INFO] Step 37100/57800; acc:  68.44; ppl:  3.68; xent: 1.30; lr: 0.00046; 6762/6629 tok/s;   9073 sec\n",
            "[2020-10-23 20:29:46,808 INFO] Step 37150/57800; acc:  68.67; ppl:  3.64; xent: 1.29; lr: 0.00046; 6565/6686 tok/s;   9127 sec\n",
            "[2020-10-23 20:30:41,124 INFO] Step 37200/57800; acc:  69.36; ppl:  3.52; xent: 1.26; lr: 0.00046; 6774/6661 tok/s;   9181 sec\n",
            "[2020-10-23 20:31:35,208 INFO] Step 37250/57800; acc:  68.72; ppl:  3.60; xent: 1.28; lr: 0.00046; 6607/6627 tok/s;   9236 sec\n",
            "[2020-10-23 20:32:29,427 INFO] Step 37300/57800; acc:  69.73; ppl:  3.44; xent: 1.23; lr: 0.00046; 6631/6755 tok/s;   9290 sec\n",
            "[2020-10-23 20:33:22,813 INFO] Step 37350/57800; acc:  68.89; ppl:  3.57; xent: 1.27; lr: 0.00046; 6685/6529 tok/s;   9343 sec\n",
            "[2020-10-23 20:34:16,568 INFO] Step 37400/57800; acc:  69.93; ppl:  3.41; xent: 1.23; lr: 0.00046; 6604/6623 tok/s;   9397 sec\n",
            "[2020-10-23 20:35:11,222 INFO] Step 37450/57800; acc:  69.64; ppl:  3.44; xent: 1.24; lr: 0.00046; 6547/6663 tok/s;   9452 sec\n",
            "[2020-10-23 20:36:05,592 INFO] Step 37500/57800; acc:  69.35; ppl:  3.48; xent: 1.25; lr: 0.00046; 6624/6552 tok/s;   9506 sec\n",
            "[2020-10-23 20:37:00,314 INFO] Step 37550/57800; acc:  69.59; ppl:  3.43; xent: 1.23; lr: 0.00046; 6661/6705 tok/s;   9561 sec\n",
            "[2020-10-23 20:37:54,345 INFO] Step 37600/57800; acc:  69.61; ppl:  3.43; xent: 1.23; lr: 0.00046; 6735/6678 tok/s;   9615 sec\n",
            "[2020-10-23 20:38:48,172 INFO] Step 37650/57800; acc:  70.08; ppl:  3.35; xent: 1.21; lr: 0.00046; 6669/6766 tok/s;   9669 sec\n",
            "[2020-10-23 20:39:42,292 INFO] Step 37700/57800; acc:  69.95; ppl:  3.37; xent: 1.21; lr: 0.00046; 6738/6662 tok/s;   9723 sec\n",
            "[2020-10-23 20:40:36,931 INFO] Step 37750/57800; acc:  69.78; ppl:  3.39; xent: 1.22; lr: 0.00045; 6663/6634 tok/s;   9777 sec\n",
            "[2020-10-23 20:41:31,066 INFO] Step 37800/57800; acc:  69.81; ppl:  3.38; xent: 1.22; lr: 0.00045; 6770/6682 tok/s;   9831 sec\n",
            "[2020-10-23 20:42:24,625 INFO] Step 37850/57800; acc:  70.40; ppl:  3.28; xent: 1.19; lr: 0.00045; 6691/6738 tok/s;   9885 sec\n",
            "[2020-10-23 20:43:18,875 INFO] Step 37900/57800; acc:  70.00; ppl:  3.36; xent: 1.21; lr: 0.00045; 6629/6667 tok/s;   9939 sec\n",
            "[2020-10-23 20:43:56,811 INFO] Loading dataset from data/multi-europarl.train.6.pt\n",
            "[2020-10-23 20:44:16,935 INFO] number of examples: 425325\n",
            "[2020-10-23 20:44:37,552 INFO] Step 37950/57800; acc:  69.19; ppl:  3.52; xent: 1.26; lr: 0.00045; 4506/4633 tok/s;  10018 sec\n",
            "[2020-10-23 20:45:31,368 INFO] Step 38000/57800; acc:  66.96; ppl:  4.02; xent: 1.39; lr: 0.00045; 6622/6631 tok/s;  10072 sec\n",
            "[2020-10-23 20:46:24,960 INFO] Step 38050/57800; acc:  66.98; ppl:  4.03; xent: 1.39; lr: 0.00045; 6769/6512 tok/s;  10125 sec\n",
            "[2020-10-23 20:47:18,510 INFO] Step 38100/57800; acc:  66.89; ppl:  4.03; xent: 1.39; lr: 0.00045; 6720/6631 tok/s;  10179 sec\n",
            "[2020-10-23 20:48:12,199 INFO] Step 38150/57800; acc:  67.59; ppl:  3.90; xent: 1.36; lr: 0.00045; 6780/6683 tok/s;  10233 sec\n",
            "[2020-10-23 20:49:05,961 INFO] Step 38200/57800; acc:  67.52; ppl:  3.89; xent: 1.36; lr: 0.00045; 6733/6762 tok/s;  10286 sec\n",
            "[2020-10-23 20:49:58,982 INFO] Step 38250/57800; acc:  67.79; ppl:  3.83; xent: 1.34; lr: 0.00045; 6889/6672 tok/s;  10339 sec\n",
            "[2020-10-23 20:50:54,079 INFO] Step 38300/57800; acc:  67.24; ppl:  3.91; xent: 1.36; lr: 0.00045; 6447/6679 tok/s;  10394 sec\n",
            "[2020-10-23 20:51:48,945 INFO] Step 38350/57800; acc:  67.90; ppl:  3.80; xent: 1.34; lr: 0.00045; 6563/6668 tok/s;  10449 sec\n",
            "[2020-10-23 20:52:43,060 INFO] Step 38400/57800; acc:  68.03; ppl:  3.79; xent: 1.33; lr: 0.00045; 6625/6523 tok/s;  10503 sec\n",
            "[2020-10-23 20:53:36,894 INFO] Step 38450/57800; acc:  68.65; ppl:  3.66; xent: 1.30; lr: 0.00045; 6665/6647 tok/s;  10557 sec\n",
            "[2020-10-23 20:54:30,710 INFO] Step 38500/57800; acc:  68.27; ppl:  3.73; xent: 1.32; lr: 0.00045; 6691/6711 tok/s;  10611 sec\n",
            "[2020-10-23 20:55:25,319 INFO] Step 38550/57800; acc:  68.66; ppl:  3.65; xent: 1.29; lr: 0.00045; 6605/6781 tok/s;  10666 sec\n",
            "[2020-10-23 20:56:19,459 INFO] Step 38600/57800; acc:  68.42; ppl:  3.67; xent: 1.30; lr: 0.00045; 6609/6653 tok/s;  10720 sec\n",
            "[2020-10-23 20:57:13,237 INFO] Step 38650/57800; acc:  69.47; ppl:  3.52; xent: 1.26; lr: 0.00045; 6604/6737 tok/s;  10774 sec\n",
            "[2020-10-23 20:58:07,494 INFO] Step 38700/57800; acc:  68.98; ppl:  3.57; xent: 1.27; lr: 0.00045; 6684/6721 tok/s;  10828 sec\n",
            "[2020-10-23 20:59:02,298 INFO] Step 38750/57800; acc:  68.98; ppl:  3.58; xent: 1.27; lr: 0.00045; 6614/6708 tok/s;  10883 sec\n",
            "[2020-10-23 20:59:57,697 INFO] Step 38800/57800; acc:  69.20; ppl:  3.53; xent: 1.26; lr: 0.00045; 6460/6681 tok/s;  10938 sec\n",
            "[2020-10-23 21:00:51,993 INFO] Step 38850/57800; acc:  69.26; ppl:  3.52; xent: 1.26; lr: 0.00045; 6700/6676 tok/s;  10992 sec\n",
            "[2020-10-23 21:01:46,072 INFO] Step 38900/57800; acc:  69.63; ppl:  3.46; xent: 1.24; lr: 0.00045; 6669/6756 tok/s;  11046 sec\n",
            "[2020-10-23 21:02:40,849 INFO] Step 38950/57800; acc:  69.36; ppl:  3.50; xent: 1.25; lr: 0.00045; 6634/6568 tok/s;  11101 sec\n",
            "[2020-10-23 21:03:35,141 INFO] Step 39000/57800; acc:  69.46; ppl:  3.47; xent: 1.24; lr: 0.00045; 6755/6649 tok/s;  11155 sec\n",
            "[2020-10-23 21:04:29,349 INFO] Step 39050/57800; acc:  69.57; ppl:  3.43; xent: 1.23; lr: 0.00045; 6722/6595 tok/s;  11210 sec\n",
            "[2020-10-23 21:05:23,171 INFO] Step 39100/57800; acc:  69.37; ppl:  3.46; xent: 1.24; lr: 0.00045; 6729/6662 tok/s;  11264 sec\n",
            "[2020-10-23 21:06:17,597 INFO] Step 39150/57800; acc:  70.43; ppl:  3.30; xent: 1.19; lr: 0.00045; 6613/6592 tok/s;  11318 sec\n",
            "[2020-10-23 21:07:11,642 INFO] Step 39200/57800; acc:  70.16; ppl:  3.34; xent: 1.21; lr: 0.00045; 6719/6729 tok/s;  11372 sec\n",
            "[2020-10-23 21:08:06,471 INFO] Step 39250/57800; acc:  70.17; ppl:  3.32; xent: 1.20; lr: 0.00045; 6573/6684 tok/s;  11427 sec\n",
            "[2020-10-23 21:09:00,956 INFO] Step 39300/57800; acc:  70.17; ppl:  3.32; xent: 1.20; lr: 0.00045; 6603/6715 tok/s;  11481 sec\n",
            "[2020-10-23 21:09:55,356 INFO] Step 39350/57800; acc:  69.76; ppl:  3.37; xent: 1.22; lr: 0.00045; 6689/6666 tok/s;  11536 sec\n",
            "[2020-10-23 21:10:50,040 INFO] Step 39400/57800; acc:  70.42; ppl:  3.28; xent: 1.19; lr: 0.00045; 6682/6693 tok/s;  11590 sec\n",
            "[2020-10-23 21:11:32,601 INFO] Loading dataset from data/multi-europarl.train.7.pt\n",
            "[2020-10-23 21:11:54,388 INFO] number of examples: 426706\n",
            "[2020-10-23 21:12:10,998 INFO] Step 39450/57800; acc:  69.58; ppl:  3.44; xent: 1.24; lr: 0.00045; 4453/4528 tok/s;  11671 sec\n",
            "[2020-10-23 21:13:05,234 INFO] Step 39500/57800; acc:  66.76; ppl:  4.10; xent: 1.41; lr: 0.00044; 6561/6649 tok/s;  11726 sec\n",
            "[2020-10-23 21:13:59,433 INFO] Step 39550/57800; acc:  67.47; ppl:  3.92; xent: 1.37; lr: 0.00044; 6683/6770 tok/s;  11780 sec\n",
            "[2020-10-23 21:14:54,369 INFO] Step 39600/57800; acc:  67.31; ppl:  3.97; xent: 1.38; lr: 0.00044; 6654/6615 tok/s;  11835 sec\n",
            "[2020-10-23 21:15:48,057 INFO] Step 39650/57800; acc:  67.66; ppl:  3.87; xent: 1.35; lr: 0.00044; 6779/6727 tok/s;  11888 sec\n",
            "[2020-10-23 21:16:41,893 INFO] Step 39700/57800; acc:  68.16; ppl:  3.77; xent: 1.33; lr: 0.00044; 6705/6714 tok/s;  11942 sec\n",
            "[2020-10-23 21:17:36,265 INFO] Step 39750/57800; acc:  67.95; ppl:  3.80; xent: 1.33; lr: 0.00044; 6723/6632 tok/s;  11997 sec\n",
            "[2020-10-23 21:18:30,548 INFO] Step 39800/57800; acc:  67.89; ppl:  3.80; xent: 1.34; lr: 0.00044; 6590/6597 tok/s;  12051 sec\n",
            "[2020-10-23 21:19:25,467 INFO] Step 39850/57800; acc:  68.01; ppl:  3.77; xent: 1.33; lr: 0.00044; 6439/6710 tok/s;  12106 sec\n",
            "[2020-10-23 21:20:19,391 INFO] Step 39900/57800; acc:  68.54; ppl:  3.68; xent: 1.30; lr: 0.00044; 6697/6653 tok/s;  12160 sec\n",
            "[2020-10-23 21:21:13,813 INFO] Step 39950/57800; acc:  68.03; ppl:  3.76; xent: 1.32; lr: 0.00044; 6612/6697 tok/s;  12214 sec\n",
            "[2020-10-23 21:22:07,692 INFO] Step 40000/57800; acc:  68.47; ppl:  3.68; xent: 1.30; lr: 0.00044; 6580/6580 tok/s;  12268 sec\n",
            "[2020-10-23 21:23:01,936 INFO] Step 40050/57800; acc:  68.83; ppl:  3.61; xent: 1.28; lr: 0.00044; 6717/6721 tok/s;  12322 sec\n",
            "[2020-10-23 21:23:56,650 INFO] Step 40100/57800; acc:  68.82; ppl:  3.62; xent: 1.29; lr: 0.00044; 6625/6670 tok/s;  12377 sec\n",
            "[2020-10-23 21:24:50,457 INFO] Step 40150/57800; acc:  69.00; ppl:  3.55; xent: 1.27; lr: 0.00044; 6654/6632 tok/s;  12431 sec\n",
            "[2020-10-23 21:25:44,433 INFO] Step 40200/57800; acc:  69.40; ppl:  3.50; xent: 1.25; lr: 0.00044; 6504/6755 tok/s;  12485 sec\n",
            "[2020-10-23 21:26:39,131 INFO] Step 40250/57800; acc:  69.07; ppl:  3.56; xent: 1.27; lr: 0.00044; 6711/6634 tok/s;  12539 sec\n",
            "[2020-10-23 21:27:33,339 INFO] Step 40300/57800; acc:  69.14; ppl:  3.53; xent: 1.26; lr: 0.00044; 6530/6558 tok/s;  12594 sec\n",
            "[2020-10-23 21:28:27,377 INFO] Step 40350/57800; acc:  69.70; ppl:  3.44; xent: 1.24; lr: 0.00044; 6810/6712 tok/s;  12648 sec\n",
            "[2020-10-23 21:29:21,801 INFO] Step 40400/57800; acc:  69.61; ppl:  3.45; xent: 1.24; lr: 0.00044; 6797/6657 tok/s;  12702 sec\n",
            "[2020-10-23 21:30:16,190 INFO] Step 40450/57800; acc:  69.93; ppl:  3.39; xent: 1.22; lr: 0.00044; 6562/6741 tok/s;  12757 sec\n",
            "[2020-10-23 21:31:09,551 INFO] Step 40500/57800; acc:  70.14; ppl:  3.36; xent: 1.21; lr: 0.00044; 6787/6697 tok/s;  12810 sec\n",
            "[2020-10-23 21:32:03,429 INFO] Step 40550/57800; acc:  69.61; ppl:  3.44; xent: 1.24; lr: 0.00044; 6724/6699 tok/s;  12864 sec\n",
            "[2020-10-23 21:32:58,085 INFO] Step 40600/57800; acc:  70.16; ppl:  3.35; xent: 1.21; lr: 0.00044; 6693/6677 tok/s;  12918 sec\n",
            "[2020-10-23 21:33:52,352 INFO] Step 40650/57800; acc:  70.02; ppl:  3.34; xent: 1.21; lr: 0.00044; 6616/6712 tok/s;  12973 sec\n",
            "[2020-10-23 21:34:46,602 INFO] Step 40700/57800; acc:  70.47; ppl:  3.29; xent: 1.19; lr: 0.00044; 6720/6717 tok/s;  13027 sec\n",
            "[2020-10-23 21:35:40,421 INFO] Step 40750/57800; acc:  70.03; ppl:  3.34; xent: 1.21; lr: 0.00044; 6649/6640 tok/s;  13081 sec\n",
            "[2020-10-23 21:36:34,958 INFO] Step 40800/57800; acc:  70.44; ppl:  3.29; xent: 1.19; lr: 0.00044; 6744/6667 tok/s;  13135 sec\n",
            "[2020-10-23 21:37:28,206 INFO] Step 40850/57800; acc:  71.05; ppl:  3.18; xent: 1.16; lr: 0.00044; 6777/6717 tok/s;  13189 sec\n",
            "[2020-10-23 21:38:23,047 INFO] Step 40900/57800; acc:  70.59; ppl:  3.26; xent: 1.18; lr: 0.00044; 6588/6700 tok/s;  13243 sec\n",
            "[2020-10-23 21:39:13,655 INFO] Loading dataset from data/multi-europarl.train.8.pt\n",
            "[2020-10-23 21:39:35,343 INFO] number of examples: 427315\n",
            "[2020-10-23 21:39:42,126 INFO] Step 40950/57800; acc:  70.63; ppl:  3.23; xent: 1.17; lr: 0.00044; 4609/4496 tok/s;  13322 sec\n",
            "[2020-10-23 21:40:36,526 INFO] Step 41000/57800; acc:  66.99; ppl:  4.02; xent: 1.39; lr: 0.00044; 6639/6665 tok/s;  13377 sec\n",
            "[2020-10-23 21:41:30,319 INFO] Step 41050/57800; acc:  67.26; ppl:  3.97; xent: 1.38; lr: 0.00044; 6680/6622 tok/s;  13431 sec\n",
            "[2020-10-23 21:42:24,297 INFO] Step 41100/57800; acc:  67.38; ppl:  3.92; xent: 1.37; lr: 0.00044; 6674/6711 tok/s;  13485 sec\n",
            "[2020-10-23 21:43:18,703 INFO] Step 41150/57800; acc:  67.69; ppl:  3.86; xent: 1.35; lr: 0.00044; 6504/6652 tok/s;  13539 sec\n",
            "[2020-10-23 21:44:13,130 INFO] Step 41200/57800; acc:  68.06; ppl:  3.79; xent: 1.33; lr: 0.00044; 6655/6689 tok/s;  13593 sec\n",
            "[2020-10-23 21:45:07,266 INFO] Step 41250/57800; acc:  68.36; ppl:  3.75; xent: 1.32; lr: 0.00044; 6704/6628 tok/s;  13648 sec\n",
            "[2020-10-23 21:46:01,648 INFO] Step 41300/57800; acc:  68.05; ppl:  3.77; xent: 1.33; lr: 0.00043; 6689/6640 tok/s;  13702 sec\n",
            "[2020-10-23 21:46:55,761 INFO] Step 41350/57800; acc:  68.19; ppl:  3.73; xent: 1.32; lr: 0.00043; 6590/6666 tok/s;  13756 sec\n",
            "[2020-10-23 21:47:50,095 INFO] Step 41400/57800; acc:  68.54; ppl:  3.67; xent: 1.30; lr: 0.00043; 6746/6727 tok/s;  13810 sec\n",
            "[2020-10-23 21:48:44,810 INFO] Step 41450/57800; acc:  68.94; ppl:  3.60; xent: 1.28; lr: 0.00043; 6646/6709 tok/s;  13865 sec\n",
            "[2020-10-23 21:49:39,713 INFO] Step 41500/57800; acc:  68.77; ppl:  3.62; xent: 1.29; lr: 0.00043; 6519/6784 tok/s;  13920 sec\n",
            "[2020-10-23 21:50:33,986 INFO] Step 41550/57800; acc:  69.15; ppl:  3.56; xent: 1.27; lr: 0.00043; 6609/6597 tok/s;  13974 sec\n",
            "[2020-10-23 21:51:28,059 INFO] Step 41600/57800; acc:  69.27; ppl:  3.53; xent: 1.26; lr: 0.00043; 6696/6738 tok/s;  14028 sec\n",
            "[2020-10-23 21:52:22,349 INFO] Step 41650/57800; acc:  69.18; ppl:  3.52; xent: 1.26; lr: 0.00043; 6638/6737 tok/s;  14083 sec\n",
            "[2020-10-23 21:53:15,757 INFO] Step 41700/57800; acc:  69.90; ppl:  3.43; xent: 1.23; lr: 0.00043; 6846/6673 tok/s;  14136 sec\n",
            "[2020-10-23 21:54:09,541 INFO] Step 41750/57800; acc:  69.90; ppl:  3.43; xent: 1.23; lr: 0.00043; 6673/6631 tok/s;  14190 sec\n",
            "[2020-10-23 21:55:03,979 INFO] Step 41800/57800; acc:  69.75; ppl:  3.43; xent: 1.23; lr: 0.00043; 6644/6725 tok/s;  14244 sec\n",
            "[2020-10-23 21:55:58,118 INFO] Step 41850/57800; acc:  69.18; ppl:  3.51; xent: 1.26; lr: 0.00043; 6714/6651 tok/s;  14298 sec\n",
            "[2020-10-23 21:56:52,417 INFO] Step 41900/57800; acc:  69.43; ppl:  3.46; xent: 1.24; lr: 0.00043; 6622/6635 tok/s;  14353 sec\n",
            "[2020-10-23 21:57:46,471 INFO] Step 41950/57800; acc:  70.00; ppl:  3.37; xent: 1.21; lr: 0.00043; 6827/6670 tok/s;  14407 sec\n",
            "[2020-10-23 21:58:40,374 INFO] Step 42000/57800; acc:  70.01; ppl:  3.37; xent: 1.22; lr: 0.00043; 6753/6662 tok/s;  14461 sec\n",
            "[2020-10-23 21:59:34,353 INFO] Step 42050/57800; acc:  70.04; ppl:  3.37; xent: 1.22; lr: 0.00043; 6810/6645 tok/s;  14515 sec\n",
            "[2020-10-23 22:00:28,956 INFO] Step 42100/57800; acc:  70.12; ppl:  3.34; xent: 1.20; lr: 0.00043; 6562/6692 tok/s;  14569 sec\n",
            "[2020-10-23 22:01:22,424 INFO] Step 42150/57800; acc:  70.52; ppl:  3.27; xent: 1.19; lr: 0.00043; 6719/6601 tok/s;  14623 sec\n",
            "[2020-10-23 22:02:15,947 INFO] Step 42200/57800; acc:  70.14; ppl:  3.34; xent: 1.21; lr: 0.00043; 6730/6640 tok/s;  14676 sec\n",
            "[2020-10-23 22:03:10,851 INFO] Step 42250/57800; acc:  70.01; ppl:  3.35; xent: 1.21; lr: 0.00043; 6572/6630 tok/s;  14731 sec\n",
            "[2020-10-23 22:04:05,486 INFO] Step 42300/57800; acc:  70.27; ppl:  3.30; xent: 1.19; lr: 0.00043; 6547/6642 tok/s;  14786 sec\n",
            "[2020-10-23 22:04:59,722 INFO] Step 42350/57800; acc:  70.62; ppl:  3.24; xent: 1.17; lr: 0.00043; 6606/6661 tok/s;  14840 sec\n",
            "[2020-10-23 22:05:54,078 INFO] Step 42400/57800; acc:  70.30; ppl:  3.28; xent: 1.19; lr: 0.00043; 6588/6677 tok/s;  14894 sec\n",
            "[2020-10-23 22:06:51,707 INFO] Step 42450/57800; acc:  71.03; ppl:  3.20; xent: 1.16; lr: 0.00043; 6286/6339 tok/s;  14952 sec\n",
            "[2020-10-23 22:06:58,378 INFO] Loading dataset from data/multi-europarl.train.9.pt\n",
            "[2020-10-23 22:07:27,585 INFO] number of examples: 425918\n",
            "[2020-10-23 22:08:19,845 INFO] Step 42500/57800; acc:  67.79; ppl:  3.84; xent: 1.35; lr: 0.00043; 4110/4101 tok/s;  15040 sec\n",
            "[2020-10-23 22:09:14,570 INFO] Step 42550/57800; acc:  67.32; ppl:  3.96; xent: 1.38; lr: 0.00043; 6525/6672 tok/s;  15095 sec\n",
            "[2020-10-23 22:10:09,061 INFO] Step 42600/57800; acc:  67.93; ppl:  3.80; xent: 1.34; lr: 0.00043; 6578/6758 tok/s;  15149 sec\n",
            "[2020-10-23 22:11:04,407 INFO] Step 42650/57800; acc:  67.80; ppl:  3.81; xent: 1.34; lr: 0.00043; 6529/6745 tok/s;  15205 sec\n",
            "[2020-10-23 22:11:59,464 INFO] Step 42700/57800; acc:  67.86; ppl:  3.80; xent: 1.34; lr: 0.00043; 6465/6669 tok/s;  15260 sec\n",
            "[2020-10-23 22:12:52,873 INFO] Step 42750/57800; acc:  68.29; ppl:  3.74; xent: 1.32; lr: 0.00043; 6866/6649 tok/s;  15313 sec\n",
            "[2020-10-23 22:13:47,583 INFO] Step 42800/57800; acc:  68.55; ppl:  3.68; xent: 1.30; lr: 0.00043; 6540/6773 tok/s;  15368 sec\n",
            "[2020-10-23 22:14:41,740 INFO] Step 42850/57800; acc:  68.63; ppl:  3.66; xent: 1.30; lr: 0.00043; 6707/6699 tok/s;  15422 sec\n",
            "[2020-10-23 22:15:35,807 INFO] Step 42900/57800; acc:  68.30; ppl:  3.70; xent: 1.31; lr: 0.00043; 6720/6612 tok/s;  15476 sec\n",
            "[2020-10-23 22:16:29,695 INFO] Step 42950/57800; acc:  68.62; ppl:  3.66; xent: 1.30; lr: 0.00043; 6759/6625 tok/s;  15530 sec\n",
            "[2020-10-23 22:17:23,922 INFO] Step 43000/57800; acc:  68.97; ppl:  3.58; xent: 1.28; lr: 0.00043; 6717/6502 tok/s;  15584 sec\n",
            "[2020-10-23 22:18:17,473 INFO] Step 43050/57800; acc:  68.89; ppl:  3.60; xent: 1.28; lr: 0.00043; 6779/6687 tok/s;  15638 sec\n",
            "[2020-10-23 22:19:12,743 INFO] Step 43100/57800; acc:  68.98; ppl:  3.57; xent: 1.27; lr: 0.00043; 6538/6656 tok/s;  15693 sec\n",
            "[2020-10-23 22:20:07,235 INFO] Step 43150/57800; acc:  69.14; ppl:  3.53; xent: 1.26; lr: 0.00043; 6654/6690 tok/s;  15748 sec\n",
            "[2020-10-23 22:21:01,199 INFO] Step 43200/57800; acc:  69.27; ppl:  3.51; xent: 1.26; lr: 0.00043; 6692/6635 tok/s;  15802 sec\n",
            "[2020-10-23 22:21:54,852 INFO] Step 43250/57800; acc:  69.38; ppl:  3.47; xent: 1.25; lr: 0.00043; 6677/6701 tok/s;  15855 sec\n",
            "[2020-10-23 22:22:48,944 INFO] Step 43300/57800; acc:  69.90; ppl:  3.41; xent: 1.23; lr: 0.00042; 6773/6662 tok/s;  15909 sec\n",
            "[2020-10-23 22:23:41,934 INFO] Step 43350/57800; acc:  69.86; ppl:  3.40; xent: 1.22; lr: 0.00042; 6775/6648 tok/s;  15962 sec\n",
            "[2020-10-23 22:23:42,657 INFO] Saving checkpoint models/multi-europarl-v3_step_43350.pt\n",
            "[2020-10-23 22:25:09,827 INFO] Step 43400/57800; acc:  69.71; ppl:  3.43; xent: 1.23; lr: 0.00042; 4007/4083 tok/s;  16050 sec\n",
            "[2020-10-23 22:26:03,955 INFO] Step 43450/57800; acc:  70.10; ppl:  3.36; xent: 1.21; lr: 0.00042; 6693/6626 tok/s;  16104 sec\n",
            "[2020-10-23 22:26:58,011 INFO] Step 43500/57800; acc:  70.28; ppl:  3.32; xent: 1.20; lr: 0.00042; 6674/6783 tok/s;  16158 sec\n",
            "[2020-10-23 22:27:52,861 INFO] Step 43550/57800; acc:  69.93; ppl:  3.36; xent: 1.21; lr: 0.00042; 6578/6614 tok/s;  16213 sec\n",
            "[2020-10-23 22:28:46,974 INFO] Step 43600/57800; acc:  70.23; ppl:  3.32; xent: 1.20; lr: 0.00042; 6700/6697 tok/s;  16267 sec\n",
            "[2020-10-23 22:29:41,032 INFO] Step 43650/57800; acc:  70.43; ppl:  3.28; xent: 1.19; lr: 0.00042; 6831/6689 tok/s;  16321 sec\n",
            "[2020-10-23 22:30:35,397 INFO] Step 43700/57800; acc:  70.50; ppl:  3.28; xent: 1.19; lr: 0.00042; 6585/6719 tok/s;  16376 sec\n",
            "[2020-10-23 22:31:28,468 INFO] Step 43750/57800; acc:  70.11; ppl:  3.32; xent: 1.20; lr: 0.00042; 6761/6528 tok/s;  16429 sec\n",
            "[2020-10-23 22:32:23,144 INFO] Step 43800/57800; acc:  70.56; ppl:  3.25; xent: 1.18; lr: 0.00042; 6610/6768 tok/s;  16483 sec\n",
            "[2020-10-23 22:33:18,080 INFO] Step 43850/57800; acc:  70.80; ppl:  3.21; xent: 1.17; lr: 0.00042; 6526/6689 tok/s;  16538 sec\n",
            "[2020-10-23 22:34:12,138 INFO] Step 43900/57800; acc:  70.75; ppl:  3.22; xent: 1.17; lr: 0.00042; 6729/6603 tok/s;  16592 sec\n",
            "[2020-10-23 22:35:06,400 INFO] Step 43950/57800; acc:  70.83; ppl:  3.20; xent: 1.16; lr: 0.00042; 6632/6721 tok/s;  16647 sec\n",
            "[2020-10-23 22:35:10,437 INFO] Loading dataset from data/multi-europarl.train.10.pt\n",
            "[2020-10-23 22:35:32,618 INFO] number of examples: 426171\n",
            "[2020-10-23 22:36:27,535 INFO] Step 44000/57800; acc:  67.85; ppl:  3.86; xent: 1.35; lr: 0.00042; 4464/4500 tok/s;  16728 sec\n",
            "[2020-10-23 22:37:21,882 INFO] Step 44050/57800; acc:  67.09; ppl:  3.97; xent: 1.38; lr: 0.00042; 6620/6616 tok/s;  16782 sec\n",
            "[2020-10-23 22:38:16,598 INFO] Step 44100/57800; acc:  67.71; ppl:  3.84; xent: 1.34; lr: 0.00042; 6599/6795 tok/s;  16837 sec\n",
            "[2020-10-23 22:39:11,275 INFO] Step 44150/57800; acc:  67.75; ppl:  3.84; xent: 1.34; lr: 0.00042; 6609/6674 tok/s;  16892 sec\n",
            "[2020-10-23 22:40:05,495 INFO] Step 44200/57800; acc:  68.42; ppl:  3.72; xent: 1.31; lr: 0.00042; 6706/6725 tok/s;  16946 sec\n",
            "[2020-10-23 22:40:59,272 INFO] Step 44250/57800; acc:  68.56; ppl:  3.68; xent: 1.30; lr: 0.00042; 6678/6649 tok/s;  17000 sec\n",
            "[2020-10-23 22:41:53,715 INFO] Step 44300/57800; acc:  68.74; ppl:  3.65; xent: 1.30; lr: 0.00042; 6647/6728 tok/s;  17054 sec\n",
            "[2020-10-23 22:42:47,611 INFO] Step 44350/57800; acc:  68.69; ppl:  3.65; xent: 1.30; lr: 0.00042; 6707/6627 tok/s;  17108 sec\n",
            "[2020-10-23 22:43:41,840 INFO] Step 44400/57800; acc:  68.88; ppl:  3.59; xent: 1.28; lr: 0.00042; 6672/6693 tok/s;  17162 sec\n",
            "[2020-10-23 22:44:36,371 INFO] Step 44450/57800; acc:  68.96; ppl:  3.60; xent: 1.28; lr: 0.00042; 6625/6645 tok/s;  17217 sec\n",
            "[2020-10-23 22:45:31,400 INFO] Step 44500/57800; acc:  68.95; ppl:  3.57; xent: 1.27; lr: 0.00042; 6563/6737 tok/s;  17272 sec\n",
            "[2020-10-23 22:46:25,942 INFO] Step 44550/57800; acc:  68.62; ppl:  3.62; xent: 1.29; lr: 0.00042; 6690/6629 tok/s;  17326 sec\n",
            "[2020-10-23 22:47:19,901 INFO] Step 44600/57800; acc:  69.06; ppl:  3.55; xent: 1.27; lr: 0.00042; 6760/6597 tok/s;  17380 sec\n",
            "[2020-10-23 22:48:14,232 INFO] Step 44650/57800; acc:  69.11; ppl:  3.53; xent: 1.26; lr: 0.00042; 6610/6595 tok/s;  17435 sec\n",
            "[2020-10-23 22:49:08,512 INFO] Step 44700/57800; acc:  69.54; ppl:  3.47; xent: 1.24; lr: 0.00042; 6714/6660 tok/s;  17489 sec\n",
            "[2020-10-23 22:50:02,763 INFO] Step 44750/57800; acc:  69.65; ppl:  3.44; xent: 1.24; lr: 0.00042; 6607/6645 tok/s;  17543 sec\n",
            "[2020-10-23 22:50:56,939 INFO] Step 44800/57800; acc:  69.70; ppl:  3.41; xent: 1.23; lr: 0.00042; 6611/6641 tok/s;  17597 sec\n",
            "[2020-10-23 22:51:50,453 INFO] Step 44850/57800; acc:  70.13; ppl:  3.35; xent: 1.21; lr: 0.00042; 6693/6725 tok/s;  17651 sec\n",
            "[2020-10-23 22:52:44,557 INFO] Step 44900/57800; acc:  70.53; ppl:  3.29; xent: 1.19; lr: 0.00042; 6625/6671 tok/s;  17705 sec\n",
            "[2020-10-23 22:53:38,944 INFO] Step 44950/57800; acc:  70.06; ppl:  3.38; xent: 1.22; lr: 0.00042; 6463/6499 tok/s;  17759 sec\n",
            "[2020-10-23 22:54:32,953 INFO] Step 45000/57800; acc:  70.14; ppl:  3.32; xent: 1.20; lr: 0.00042; 6762/6707 tok/s;  17813 sec\n",
            "[2020-10-23 22:55:26,825 INFO] Step 45050/57800; acc:  70.31; ppl:  3.31; xent: 1.20; lr: 0.00042; 6819/6688 tok/s;  17867 sec\n",
            "[2020-10-23 22:56:21,447 INFO] Step 45100/57800; acc:  70.11; ppl:  3.33; xent: 1.20; lr: 0.00042; 6708/6673 tok/s;  17922 sec\n",
            "[2020-10-23 22:57:14,674 INFO] Step 45150/57800; acc:  70.99; ppl:  3.21; xent: 1.17; lr: 0.00042; 6734/6747 tok/s;  17975 sec\n",
            "[2020-10-23 22:58:09,612 INFO] Step 45200/57800; acc:  70.73; ppl:  3.24; xent: 1.18; lr: 0.00042; 6607/6699 tok/s;  18030 sec\n",
            "[2020-10-23 22:59:03,834 INFO] Step 45250/57800; acc:  71.06; ppl:  3.19; xent: 1.16; lr: 0.00042; 6738/6646 tok/s;  18084 sec\n",
            "[2020-10-23 22:59:57,838 INFO] Step 45300/57800; acc:  70.94; ppl:  3.19; xent: 1.16; lr: 0.00042; 6741/6708 tok/s;  18138 sec\n",
            "[2020-10-23 23:00:51,346 INFO] Step 45350/57800; acc:  71.04; ppl:  3.18; xent: 1.16; lr: 0.00042; 6794/6714 tok/s;  18192 sec\n",
            "[2020-10-23 23:01:44,897 INFO] Step 45400/57800; acc:  71.30; ppl:  3.16; xent: 1.15; lr: 0.00041; 6772/6687 tok/s;  18245 sec\n",
            "[2020-10-23 23:02:39,150 INFO] Step 45450/57800; acc:  71.06; ppl:  3.17; xent: 1.15; lr: 0.00041; 6501/6715 tok/s;  18299 sec\n",
            "[2020-10-23 23:02:52,373 INFO] Loading dataset from data/multi-europarl.train.11.pt\n",
            "[2020-10-23 23:03:13,236 INFO] number of examples: 426397\n",
            "[2020-10-23 23:03:59,468 INFO] Step 45500/57800; acc:  68.30; ppl:  3.73; xent: 1.32; lr: 0.00041; 4501/4494 tok/s;  18380 sec\n",
            "[2020-10-23 23:04:54,444 INFO] Step 45550/57800; acc:  67.67; ppl:  3.87; xent: 1.35; lr: 0.00041; 6594/6751 tok/s;  18435 sec\n",
            "[2020-10-23 23:05:48,351 INFO] Step 45600/57800; acc:  67.79; ppl:  3.83; xent: 1.34; lr: 0.00041; 6730/6562 tok/s;  18489 sec\n",
            "[2020-10-23 23:06:42,882 INFO] Step 45650/57800; acc:  68.19; ppl:  3.75; xent: 1.32; lr: 0.00041; 6515/6763 tok/s;  18543 sec\n",
            "[2020-10-23 23:07:36,628 INFO] Step 45700/57800; acc:  68.35; ppl:  3.73; xent: 1.32; lr: 0.00041; 6733/6710 tok/s;  18597 sec\n",
            "[2020-10-23 23:08:30,584 INFO] Step 45750/57800; acc:  68.22; ppl:  3.73; xent: 1.32; lr: 0.00041; 6677/6594 tok/s;  18651 sec\n",
            "[2020-10-23 23:09:24,718 INFO] Step 45800/57800; acc:  68.65; ppl:  3.64; xent: 1.29; lr: 0.00041; 6753/6648 tok/s;  18705 sec\n",
            "[2020-10-23 23:10:18,587 INFO] Step 45850/57800; acc:  68.50; ppl:  3.66; xent: 1.30; lr: 0.00041; 6742/6729 tok/s;  18759 sec\n",
            "[2020-10-23 23:11:13,142 INFO] Step 45900/57800; acc:  68.87; ppl:  3.60; xent: 1.28; lr: 0.00041; 6639/6694 tok/s;  18813 sec\n",
            "[2020-10-23 23:12:08,142 INFO] Step 45950/57800; acc:  68.86; ppl:  3.59; xent: 1.28; lr: 0.00041; 6397/6646 tok/s;  18868 sec\n",
            "[2020-10-23 23:13:02,711 INFO] Step 46000/57800; acc:  68.96; ppl:  3.59; xent: 1.28; lr: 0.00041; 6596/6637 tok/s;  18923 sec\n",
            "[2020-10-23 23:13:56,277 INFO] Step 46050/57800; acc:  69.42; ppl:  3.51; xent: 1.25; lr: 0.00041; 6739/6646 tok/s;  18977 sec\n",
            "[2020-10-23 23:14:51,336 INFO] Step 46100/57800; acc:  69.77; ppl:  3.44; xent: 1.24; lr: 0.00041; 6596/6687 tok/s;  19032 sec\n",
            "[2020-10-23 23:15:45,030 INFO] Step 46150/57800; acc:  69.63; ppl:  3.45; xent: 1.24; lr: 0.00041; 6716/6572 tok/s;  19085 sec\n",
            "[2020-10-23 23:16:39,132 INFO] Step 46200/57800; acc:  69.64; ppl:  3.46; xent: 1.24; lr: 0.00041; 6700/6583 tok/s;  19139 sec\n",
            "[2020-10-23 23:17:32,643 INFO] Step 46250/57800; acc:  69.70; ppl:  3.42; xent: 1.23; lr: 0.00041; 6803/6646 tok/s;  19193 sec\n",
            "[2020-10-23 23:18:27,004 INFO] Step 46300/57800; acc:  69.84; ppl:  3.40; xent: 1.22; lr: 0.00041; 6547/6683 tok/s;  19247 sec\n",
            "[2020-10-23 23:19:20,531 INFO] Step 46350/57800; acc:  70.83; ppl:  3.26; xent: 1.18; lr: 0.00041; 6633/6534 tok/s;  19301 sec\n",
            "[2020-10-23 23:20:14,625 INFO] Step 46400/57800; acc:  69.97; ppl:  3.38; xent: 1.22; lr: 0.00041; 6586/6740 tok/s;  19355 sec\n",
            "[2020-10-23 23:21:08,482 INFO] Step 46450/57800; acc:  70.28; ppl:  3.34; xent: 1.20; lr: 0.00041; 6769/6697 tok/s;  19409 sec\n",
            "[2020-10-23 23:22:02,988 INFO] Step 46500/57800; acc:  69.77; ppl:  3.40; xent: 1.22; lr: 0.00041; 6680/6620 tok/s;  19463 sec\n",
            "[2020-10-23 23:22:56,592 INFO] Step 46550/57800; acc:  70.85; ppl:  3.23; xent: 1.17; lr: 0.00041; 6695/6737 tok/s;  19517 sec\n",
            "[2020-10-23 23:23:50,821 INFO] Step 46600/57800; acc:  70.64; ppl:  3.25; xent: 1.18; lr: 0.00041; 6681/6685 tok/s;  19571 sec\n",
            "[2020-10-23 23:24:45,742 INFO] Step 46650/57800; acc:  70.73; ppl:  3.24; xent: 1.18; lr: 0.00041; 6565/6722 tok/s;  19626 sec\n",
            "[2020-10-23 23:25:38,888 INFO] Step 46700/57800; acc:  71.11; ppl:  3.19; xent: 1.16; lr: 0.00041; 6837/6735 tok/s;  19679 sec\n",
            "[2020-10-23 23:26:33,602 INFO] Step 46750/57800; acc:  71.13; ppl:  3.17; xent: 1.15; lr: 0.00041; 6561/6724 tok/s;  19734 sec\n",
            "[2020-10-23 23:27:27,462 INFO] Step 46800/57800; acc:  70.87; ppl:  3.21; xent: 1.17; lr: 0.00041; 6796/6598 tok/s;  19788 sec\n",
            "[2020-10-23 23:28:22,319 INFO] Step 46850/57800; acc:  71.25; ppl:  3.14; xent: 1.14; lr: 0.00041; 6622/6721 tok/s;  19843 sec\n",
            "[2020-10-23 23:29:16,251 INFO] Step 46900/57800; acc:  71.09; ppl:  3.16; xent: 1.15; lr: 0.00041; 6660/6737 tok/s;  19897 sec\n",
            "[2020-10-23 23:30:10,697 INFO] Step 46950/57800; acc:  71.10; ppl:  3.16; xent: 1.15; lr: 0.00041; 6673/6725 tok/s;  19951 sec\n",
            "[2020-10-23 23:30:28,980 INFO] Loading dataset from data/multi-europarl.train.12.pt\n",
            "[2020-10-23 23:30:51,312 INFO] number of examples: 424283\n",
            "[2020-10-23 23:31:31,041 INFO] Step 47000/57800; acc:  68.17; ppl:  3.73; xent: 1.32; lr: 0.00041; 4460/4412 tok/s;  20031 sec\n",
            "[2020-10-23 23:32:24,057 INFO] Step 47050/57800; acc:  67.70; ppl:  3.86; xent: 1.35; lr: 0.00041; 6884/6711 tok/s;  20084 sec\n",
            "[2020-10-23 23:33:19,019 INFO] Step 47100/57800; acc:  67.57; ppl:  3.87; xent: 1.35; lr: 0.00041; 6549/6625 tok/s;  20139 sec\n",
            "[2020-10-23 23:34:13,311 INFO] Step 47150/57800; acc:  67.80; ppl:  3.83; xent: 1.34; lr: 0.00041; 6707/6641 tok/s;  20194 sec\n",
            "[2020-10-23 23:35:08,306 INFO] Step 47200/57800; acc:  68.08; ppl:  3.75; xent: 1.32; lr: 0.00041; 6454/6643 tok/s;  20249 sec\n",
            "[2020-10-23 23:36:03,034 INFO] Step 47250/57800; acc:  68.63; ppl:  3.67; xent: 1.30; lr: 0.00041; 6551/6690 tok/s;  20303 sec\n",
            "[2020-10-23 23:36:57,299 INFO] Step 47300/57800; acc:  68.25; ppl:  3.70; xent: 1.31; lr: 0.00041; 6782/6649 tok/s;  20358 sec\n",
            "[2020-10-23 23:37:51,354 INFO] Step 47350/57800; acc:  69.04; ppl:  3.56; xent: 1.27; lr: 0.00041; 6648/6765 tok/s;  20412 sec\n",
            "[2020-10-23 23:38:45,054 INFO] Step 47400/57800; acc:  68.58; ppl:  3.66; xent: 1.30; lr: 0.00041; 6506/6481 tok/s;  20465 sec\n",
            "[2020-10-23 23:39:38,978 INFO] Step 47450/57800; acc:  69.40; ppl:  3.50; xent: 1.25; lr: 0.00041; 6664/6700 tok/s;  20519 sec\n",
            "[2020-10-23 23:40:33,221 INFO] Step 47500/57800; acc:  69.12; ppl:  3.54; xent: 1.26; lr: 0.00041; 6698/6739 tok/s;  20574 sec\n",
            "[2020-10-23 23:41:27,164 INFO] Step 47550/57800; acc:  69.03; ppl:  3.57; xent: 1.27; lr: 0.00041; 6718/6654 tok/s;  20628 sec\n",
            "[2020-10-23 23:42:21,749 INFO] Step 47600/57800; acc:  69.61; ppl:  3.45; xent: 1.24; lr: 0.00041; 6631/6763 tok/s;  20682 sec\n",
            "[2020-10-23 23:43:16,166 INFO] Step 47650/57800; acc:  69.61; ppl:  3.45; xent: 1.24; lr: 0.00040; 6644/6707 tok/s;  20737 sec\n",
            "[2020-10-23 23:44:09,372 INFO] Step 47700/57800; acc:  68.96; ppl:  3.56; xent: 1.27; lr: 0.00040; 6768/6627 tok/s;  20790 sec\n",
            "[2020-10-23 23:45:03,459 INFO] Step 47750/57800; acc:  69.52; ppl:  3.46; xent: 1.24; lr: 0.00040; 6718/6705 tok/s;  20844 sec\n",
            "[2020-10-23 23:45:58,811 INFO] Step 47800/57800; acc:  69.87; ppl:  3.42; xent: 1.23; lr: 0.00040; 6653/6735 tok/s;  20899 sec\n",
            "[2020-10-23 23:46:53,377 INFO] Step 47850/57800; acc:  70.14; ppl:  3.35; xent: 1.21; lr: 0.00040; 6641/6686 tok/s;  20954 sec\n",
            "[2020-10-23 23:47:48,519 INFO] Step 47900/57800; acc:  70.38; ppl:  3.30; xent: 1.19; lr: 0.00040; 6534/6698 tok/s;  21009 sec\n",
            "[2020-10-23 23:48:42,008 INFO] Step 47950/57800; acc:  70.19; ppl:  3.34; xent: 1.21; lr: 0.00040; 6583/6704 tok/s;  21062 sec\n",
            "[2020-10-23 23:49:37,265 INFO] Step 48000/57800; acc:  70.25; ppl:  3.31; xent: 1.20; lr: 0.00040; 6541/6736 tok/s;  21118 sec\n",
            "[2020-10-23 23:50:31,321 INFO] Step 48050/57800; acc:  70.75; ppl:  3.24; xent: 1.17; lr: 0.00040; 6752/6771 tok/s;  21172 sec\n",
            "[2020-10-23 23:51:25,824 INFO] Step 48100/57800; acc:  70.40; ppl:  3.30; xent: 1.19; lr: 0.00040; 6545/6676 tok/s;  21226 sec\n",
            "[2020-10-23 23:52:20,166 INFO] Step 48150/57800; acc:  70.22; ppl:  3.29; xent: 1.19; lr: 0.00040; 6649/6672 tok/s;  21281 sec\n",
            "[2020-10-23 23:53:13,900 INFO] Step 48200/57800; acc:  71.33; ppl:  3.15; xent: 1.15; lr: 0.00040; 6860/6693 tok/s;  21334 sec\n",
            "[2020-10-23 23:54:07,801 INFO] Step 48250/57800; acc:  70.70; ppl:  3.24; xent: 1.18; lr: 0.00040; 6801/6643 tok/s;  21388 sec\n",
            "[2020-10-23 23:55:01,585 INFO] Step 48300/57800; acc:  70.98; ppl:  3.18; xent: 1.16; lr: 0.00040; 6910/6650 tok/s;  21442 sec\n",
            "[2020-10-23 23:55:54,739 INFO] Step 48350/57800; acc:  71.04; ppl:  3.19; xent: 1.16; lr: 0.00040; 6698/6528 tok/s;  21495 sec\n",
            "[2020-10-23 23:56:49,307 INFO] Step 48400/57800; acc:  70.83; ppl:  3.20; xent: 1.16; lr: 0.00040; 6474/6612 tok/s;  21550 sec\n",
            "[2020-10-23 23:57:43,844 INFO] Step 48450/57800; acc:  71.09; ppl:  3.16; xent: 1.15; lr: 0.00040; 6592/6634 tok/s;  21604 sec\n",
            "[2020-10-23 23:58:03,448 INFO] Loading dataset from data/multi-europarl.train.13.pt\n",
            "[2020-10-23 23:58:24,435 INFO] number of examples: 425633\n",
            "[2020-10-23 23:59:03,876 INFO] Step 48500/57800; acc:  68.98; ppl:  3.59; xent: 1.28; lr: 0.00040; 4536/4581 tok/s;  21684 sec\n",
            "[2020-10-23 23:59:57,421 INFO] Step 48550/57800; acc:  67.81; ppl:  3.85; xent: 1.35; lr: 0.00040; 6815/6558 tok/s;  21738 sec\n",
            "[2020-10-24 00:00:51,578 INFO] Step 48600/57800; acc:  68.22; ppl:  3.77; xent: 1.33; lr: 0.00040; 6662/6709 tok/s;  21792 sec\n",
            "[2020-10-24 00:01:45,499 INFO] Step 48650/57800; acc:  67.87; ppl:  3.81; xent: 1.34; lr: 0.00040; 6709/6627 tok/s;  21846 sec\n",
            "[2020-10-24 00:02:39,393 INFO] Step 48700/57800; acc:  68.39; ppl:  3.73; xent: 1.32; lr: 0.00040; 6694/6603 tok/s;  21900 sec\n",
            "[2020-10-24 00:03:33,998 INFO] Step 48750/57800; acc:  68.39; ppl:  3.69; xent: 1.31; lr: 0.00040; 6540/6767 tok/s;  21954 sec\n",
            "[2020-10-24 00:04:28,748 INFO] Step 48800/57800; acc:  68.87; ppl:  3.61; xent: 1.28; lr: 0.00040; 6567/6708 tok/s;  22009 sec\n",
            "[2020-10-24 00:05:22,786 INFO] Step 48850/57800; acc:  68.84; ppl:  3.62; xent: 1.29; lr: 0.00040; 6668/6639 tok/s;  22063 sec\n",
            "[2020-10-24 00:06:16,636 INFO] Step 48900/57800; acc:  69.43; ppl:  3.52; xent: 1.26; lr: 0.00040; 6623/6728 tok/s;  22117 sec\n",
            "[2020-10-24 00:07:11,239 INFO] Step 48950/57800; acc:  69.03; ppl:  3.57; xent: 1.27; lr: 0.00040; 6676/6615 tok/s;  22172 sec\n",
            "[2020-10-24 00:08:04,630 INFO] Step 49000/57800; acc:  69.57; ppl:  3.48; xent: 1.25; lr: 0.00040; 6751/6567 tok/s;  22225 sec\n",
            "[2020-10-24 00:08:58,341 INFO] Step 49050/57800; acc:  69.29; ppl:  3.51; xent: 1.26; lr: 0.00040; 6652/6735 tok/s;  22279 sec\n",
            "[2020-10-24 00:09:52,786 INFO] Step 49100/57800; acc:  69.62; ppl:  3.45; xent: 1.24; lr: 0.00040; 6613/6620 tok/s;  22333 sec\n",
            "[2020-10-24 00:10:47,728 INFO] Step 49150/57800; acc:  69.84; ppl:  3.41; xent: 1.23; lr: 0.00040; 6630/6707 tok/s;  22388 sec\n",
            "[2020-10-24 00:11:42,899 INFO] Step 49200/57800; acc:  69.49; ppl:  3.46; xent: 1.24; lr: 0.00040; 6523/6689 tok/s;  22443 sec\n",
            "[2020-10-24 00:12:36,748 INFO] Step 49250/57800; acc:  69.65; ppl:  3.44; xent: 1.24; lr: 0.00040; 6666/6719 tok/s;  22497 sec\n",
            "[2020-10-24 00:13:30,412 INFO] Step 49300/57800; acc:  69.62; ppl:  3.44; xent: 1.24; lr: 0.00040; 6788/6630 tok/s;  22551 sec\n",
            "[2020-10-24 00:14:24,740 INFO] Step 49350/57800; acc:  70.31; ppl:  3.32; xent: 1.20; lr: 0.00040; 6663/6718 tok/s;  22605 sec\n",
            "[2020-10-24 00:15:18,731 INFO] Step 49400/57800; acc:  70.33; ppl:  3.32; xent: 1.20; lr: 0.00040; 6808/6631 tok/s;  22659 sec\n",
            "[2020-10-24 00:16:13,782 INFO] Step 49450/57800; acc:  70.19; ppl:  3.33; xent: 1.20; lr: 0.00040; 6531/6655 tok/s;  22714 sec\n",
            "[2020-10-24 00:17:07,564 INFO] Step 49500/57800; acc:  70.40; ppl:  3.30; xent: 1.19; lr: 0.00040; 6628/6605 tok/s;  22768 sec\n",
            "[2020-10-24 00:18:01,188 INFO] Step 49550/57800; acc:  70.33; ppl:  3.30; xent: 1.19; lr: 0.00040; 6784/6527 tok/s;  22822 sec\n",
            "[2020-10-24 00:18:55,255 INFO] Step 49600/57800; acc:  70.43; ppl:  3.27; xent: 1.18; lr: 0.00040; 6669/6651 tok/s;  22876 sec\n",
            "[2020-10-24 00:19:49,494 INFO] Step 49650/57800; acc:  70.65; ppl:  3.24; xent: 1.18; lr: 0.00040; 6691/6710 tok/s;  22930 sec\n",
            "[2020-10-24 00:20:43,355 INFO] Step 49700/57800; acc:  70.59; ppl:  3.26; xent: 1.18; lr: 0.00040; 6766/6619 tok/s;  22984 sec\n",
            "[2020-10-24 00:21:37,450 INFO] Step 49750/57800; acc:  70.52; ppl:  3.27; xent: 1.19; lr: 0.00040; 6800/6613 tok/s;  23038 sec\n",
            "[2020-10-24 00:22:32,349 INFO] Step 49800/57800; acc:  71.20; ppl:  3.16; xent: 1.15; lr: 0.00040; 6425/6711 tok/s;  23093 sec\n",
            "[2020-10-24 00:23:26,302 INFO] Step 49850/57800; acc:  70.55; ppl:  3.25; xent: 1.18; lr: 0.00040; 6590/6692 tok/s;  23147 sec\n",
            "[2020-10-24 00:24:20,509 INFO] Step 49900/57800; acc:  71.34; ppl:  3.14; xent: 1.14; lr: 0.00040; 6647/6756 tok/s;  23201 sec\n",
            "[2020-10-24 00:25:16,193 INFO] Step 49950/57800; acc:  71.06; ppl:  3.16; xent: 1.15; lr: 0.00040; 6466/6752 tok/s;  23257 sec\n",
            "[2020-10-24 00:25:41,801 INFO] Loading dataset from data/multi-europarl.train.14.pt\n",
            "[2020-10-24 00:26:05,306 INFO] number of examples: 427342\n",
            "[2020-10-24 00:26:37,859 INFO] Step 50000/57800; acc:  69.61; ppl:  3.45; xent: 1.24; lr: 0.00040; 4444/4410 tok/s;  23338 sec\n",
            "[2020-10-24 00:27:31,560 INFO] Step 50050/57800; acc:  67.82; ppl:  3.83; xent: 1.34; lr: 0.00040; 6802/6686 tok/s;  23392 sec\n",
            "[2020-10-24 00:28:26,213 INFO] Step 50100/57800; acc:  68.23; ppl:  3.75; xent: 1.32; lr: 0.00039; 6531/6683 tok/s;  23447 sec\n",
            "[2020-10-24 00:29:20,180 INFO] Step 50150/57800; acc:  68.19; ppl:  3.74; xent: 1.32; lr: 0.00039; 6701/6684 tok/s;  23501 sec\n",
            "[2020-10-24 00:30:14,619 INFO] Step 50200/57800; acc:  68.32; ppl:  3.71; xent: 1.31; lr: 0.00039; 6590/6688 tok/s;  23555 sec\n",
            "[2020-10-24 00:31:08,659 INFO] Step 50250/57800; acc:  68.93; ppl:  3.60; xent: 1.28; lr: 0.00039; 6598/6656 tok/s;  23609 sec\n",
            "[2020-10-24 00:32:02,146 INFO] Step 50300/57800; acc:  68.10; ppl:  3.73; xent: 1.32; lr: 0.00039; 6737/6449 tok/s;  23662 sec\n",
            "[2020-10-24 00:32:55,757 INFO] Step 50350/57800; acc:  68.81; ppl:  3.62; xent: 1.29; lr: 0.00039; 6678/6592 tok/s;  23716 sec\n",
            "[2020-10-24 00:33:49,683 INFO] Step 50400/57800; acc:  69.35; ppl:  3.52; xent: 1.26; lr: 0.00039; 6716/6705 tok/s;  23770 sec\n",
            "[2020-10-24 00:34:44,152 INFO] Step 50450/57800; acc:  69.30; ppl:  3.52; xent: 1.26; lr: 0.00039; 6534/6707 tok/s;  23824 sec\n",
            "[2020-10-24 00:35:39,478 INFO] Step 50500/57800; acc:  69.33; ppl:  3.51; xent: 1.26; lr: 0.00039; 6469/6781 tok/s;  23880 sec\n",
            "[2020-10-24 00:36:32,469 INFO] Step 50550/57800; acc:  69.10; ppl:  3.52; xent: 1.26; lr: 0.00039; 6828/6645 tok/s;  23933 sec\n",
            "[2020-10-24 00:37:00,414 INFO] Saving checkpoint models/multi-europarl-v3_step_50575.pt\n",
            "[2020-10-24 00:38:01,028 INFO] Step 50600/57800; acc:  69.97; ppl:  3.40; xent: 1.22; lr: 0.00039; 4081/4159 tok/s;  24021 sec\n",
            "[2020-10-24 00:38:55,507 INFO] Step 50650/57800; acc:  69.25; ppl:  3.50; xent: 1.25; lr: 0.00039; 6597/6616 tok/s;  24076 sec\n",
            "[2020-10-24 00:39:49,522 INFO] Step 50700/57800; acc:  69.88; ppl:  3.40; xent: 1.22; lr: 0.00039; 6767/6716 tok/s;  24130 sec\n",
            "[2020-10-24 00:40:42,945 INFO] Step 50750/57800; acc:  70.33; ppl:  3.33; xent: 1.20; lr: 0.00039; 6669/6612 tok/s;  24183 sec\n",
            "[2020-10-24 00:41:37,294 INFO] Step 50800/57800; acc:  70.27; ppl:  3.33; xent: 1.20; lr: 0.00039; 6706/6759 tok/s;  24238 sec\n",
            "[2020-10-24 00:42:31,823 INFO] Step 50850/57800; acc:  69.66; ppl:  3.41; xent: 1.23; lr: 0.00039; 6683/6693 tok/s;  24292 sec\n",
            "[2020-10-24 00:43:25,941 INFO] Step 50900/57800; acc:  70.04; ppl:  3.36; xent: 1.21; lr: 0.00039; 6620/6604 tok/s;  24346 sec\n",
            "[2020-10-24 00:44:21,105 INFO] Step 50950/57800; acc:  70.20; ppl:  3.32; xent: 1.20; lr: 0.00039; 6553/6613 tok/s;  24401 sec\n",
            "[2020-10-24 00:45:15,889 INFO] Step 51000/57800; acc:  70.47; ppl:  3.29; xent: 1.19; lr: 0.00039; 6639/6714 tok/s;  24456 sec\n",
            "[2020-10-24 00:46:10,662 INFO] Step 51050/57800; acc:  70.25; ppl:  3.31; xent: 1.20; lr: 0.00039; 6557/6678 tok/s;  24511 sec\n",
            "[2020-10-24 00:47:05,184 INFO] Step 51100/57800; acc:  70.54; ppl:  3.27; xent: 1.19; lr: 0.00039; 6625/6701 tok/s;  24566 sec\n",
            "[2020-10-24 00:47:58,746 INFO] Step 51150/57800; acc:  70.67; ppl:  3.24; xent: 1.18; lr: 0.00039; 6621/6686 tok/s;  24619 sec\n",
            "[2020-10-24 00:48:52,626 INFO] Step 51200/57800; acc:  71.08; ppl:  3.20; xent: 1.16; lr: 0.00039; 6743/6652 tok/s;  24673 sec\n",
            "[2020-10-24 00:49:46,093 INFO] Step 51250/57800; acc:  70.38; ppl:  3.27; xent: 1.18; lr: 0.00039; 6839/6623 tok/s;  24726 sec\n",
            "[2020-10-24 00:50:39,867 INFO] Step 51300/57800; acc:  70.76; ppl:  3.22; xent: 1.17; lr: 0.00039; 6756/6674 tok/s;  24780 sec\n",
            "[2020-10-24 00:51:34,437 INFO] Step 51350/57800; acc:  70.99; ppl:  3.18; xent: 1.16; lr: 0.00039; 6667/6719 tok/s;  24835 sec\n",
            "[2020-10-24 00:52:29,189 INFO] Step 51400/57800; acc:  71.27; ppl:  3.13; xent: 1.14; lr: 0.00039; 6643/6705 tok/s;  24890 sec\n",
            "[2020-10-24 00:53:23,195 INFO] Step 51450/57800; acc:  71.30; ppl:  3.13; xent: 1.14; lr: 0.00039; 6714/6708 tok/s;  24944 sec\n",
            "[2020-10-24 00:53:58,005 INFO] Loading dataset from data/multi-europarl.train.15.pt\n",
            "[2020-10-24 00:54:21,624 INFO] number of examples: 427386\n",
            "[2020-10-24 00:54:45,526 INFO] Step 51500/57800; acc:  70.18; ppl:  3.35; xent: 1.21; lr: 0.00039; 4418/4380 tok/s;  25026 sec\n",
            "[2020-10-24 00:55:39,817 INFO] Step 51550/57800; acc:  68.00; ppl:  3.79; xent: 1.33; lr: 0.00039; 6675/6604 tok/s;  25080 sec\n",
            "[2020-10-24 00:56:33,992 INFO] Step 51600/57800; acc:  68.22; ppl:  3.74; xent: 1.32; lr: 0.00039; 6613/6588 tok/s;  25134 sec\n",
            "[2020-10-24 00:57:27,648 INFO] Step 51650/57800; acc:  68.43; ppl:  3.69; xent: 1.31; lr: 0.00039; 6796/6662 tok/s;  25188 sec\n",
            "[2020-10-24 00:58:22,569 INFO] Step 51700/57800; acc:  67.94; ppl:  3.77; xent: 1.33; lr: 0.00039; 6447/6633 tok/s;  25243 sec\n",
            "[2020-10-24 00:59:17,272 INFO] Step 51750/57800; acc:  69.12; ppl:  3.58; xent: 1.28; lr: 0.00039; 6568/6628 tok/s;  25298 sec\n",
            "[2020-10-24 01:00:11,027 INFO] Step 51800/57800; acc:  68.90; ppl:  3.59; xent: 1.28; lr: 0.00039; 6842/6619 tok/s;  25351 sec\n",
            "[2020-10-24 01:01:04,701 INFO] Step 51850/57800; acc:  69.08; ppl:  3.56; xent: 1.27; lr: 0.00039; 6705/6676 tok/s;  25405 sec\n",
            "[2020-10-24 01:01:59,221 INFO] Step 51900/57800; acc:  69.66; ppl:  3.47; xent: 1.24; lr: 0.00039; 6571/6685 tok/s;  25460 sec\n",
            "[2020-10-24 01:02:53,201 INFO] Step 51950/57800; acc:  69.43; ppl:  3.51; xent: 1.26; lr: 0.00039; 6720/6749 tok/s;  25514 sec\n",
            "[2020-10-24 01:03:47,467 INFO] Step 52000/57800; acc:  69.05; ppl:  3.54; xent: 1.26; lr: 0.00039; 6408/6718 tok/s;  25568 sec\n",
            "[2020-10-24 01:04:41,570 INFO] Step 52050/57800; acc:  69.42; ppl:  3.48; xent: 1.25; lr: 0.00039; 6715/6678 tok/s;  25622 sec\n",
            "[2020-10-24 01:05:36,247 INFO] Step 52100/57800; acc:  69.96; ppl:  3.39; xent: 1.22; lr: 0.00039; 6624/6716 tok/s;  25677 sec\n",
            "[2020-10-24 01:06:30,740 INFO] Step 52150/57800; acc:  69.56; ppl:  3.44; xent: 1.24; lr: 0.00039; 6656/6637 tok/s;  25731 sec\n",
            "[2020-10-24 01:07:25,296 INFO] Step 52200/57800; acc:  70.00; ppl:  3.39; xent: 1.22; lr: 0.00039; 6566/6686 tok/s;  25786 sec\n",
            "[2020-10-24 01:08:19,256 INFO] Step 52250/57800; acc:  70.38; ppl:  3.32; xent: 1.20; lr: 0.00039; 6718/6667 tok/s;  25840 sec\n",
            "[2020-10-24 01:09:14,121 INFO] Step 52300/57800; acc:  70.32; ppl:  3.31; xent: 1.20; lr: 0.00039; 6545/6728 tok/s;  25894 sec\n",
            "[2020-10-24 01:10:08,448 INFO] Step 52350/57800; acc:  70.06; ppl:  3.35; xent: 1.21; lr: 0.00039; 6769/6701 tok/s;  25949 sec\n",
            "[2020-10-24 01:11:02,472 INFO] Step 52400/57800; acc:  70.08; ppl:  3.35; xent: 1.21; lr: 0.00039; 6700/6638 tok/s;  26003 sec\n",
            "[2020-10-24 01:11:57,105 INFO] Step 52450/57800; acc:  70.02; ppl:  3.34; xent: 1.21; lr: 0.00039; 6648/6594 tok/s;  26057 sec\n",
            "[2020-10-24 01:12:50,989 INFO] Step 52500/57800; acc:  70.07; ppl:  3.34; xent: 1.21; lr: 0.00039; 6757/6728 tok/s;  26111 sec\n",
            "[2020-10-24 01:13:45,827 INFO] Step 52550/57800; acc:  70.99; ppl:  3.21; xent: 1.17; lr: 0.00039; 6546/6737 tok/s;  26166 sec\n",
            "[2020-10-24 01:14:39,480 INFO] Step 52600/57800; acc:  70.59; ppl:  3.26; xent: 1.18; lr: 0.00039; 6750/6648 tok/s;  26220 sec\n",
            "[2020-10-24 01:15:33,619 INFO] Step 52650/57800; acc:  70.60; ppl:  3.23; xent: 1.17; lr: 0.00039; 6611/6662 tok/s;  26274 sec\n",
            "[2020-10-24 01:16:27,214 INFO] Step 52700/57800; acc:  71.27; ppl:  3.16; xent: 1.15; lr: 0.00039; 6768/6634 tok/s;  26328 sec\n",
            "[2020-10-24 01:17:21,259 INFO] Step 52750/57800; acc:  71.01; ppl:  3.20; xent: 1.16; lr: 0.00038; 6716/6655 tok/s;  26382 sec\n",
            "[2020-10-24 01:18:14,950 INFO] Step 52800/57800; acc:  70.84; ppl:  3.18; xent: 1.16; lr: 0.00038; 6665/6722 tok/s;  26435 sec\n",
            "[2020-10-24 01:19:09,279 INFO] Step 52850/57800; acc:  71.04; ppl:  3.16; xent: 1.15; lr: 0.00038; 6615/6700 tok/s;  26490 sec\n",
            "[2020-10-24 01:20:03,031 INFO] Step 52900/57800; acc:  71.33; ppl:  3.13; xent: 1.14; lr: 0.00038; 6807/6717 tok/s;  26543 sec\n",
            "[2020-10-24 01:20:57,650 INFO] Step 52950/57800; acc:  71.28; ppl:  3.12; xent: 1.14; lr: 0.00038; 6619/6691 tok/s;  26598 sec\n",
            "[2020-10-24 01:21:44,590 INFO] Loading dataset from data/multi-europarl.train.16.pt\n",
            "[2020-10-24 01:22:04,616 INFO] number of examples: 425902\n",
            "[2020-10-24 01:22:16,593 INFO] Step 53000/57800; acc:  71.01; ppl:  3.18; xent: 1.16; lr: 0.00038; 4606/4600 tok/s;  26677 sec\n",
            "[2020-10-24 01:23:10,061 INFO] Step 53050/57800; acc:  68.14; ppl:  3.76; xent: 1.32; lr: 0.00038; 6786/6763 tok/s;  26730 sec\n",
            "[2020-10-24 01:24:03,438 INFO] Step 53100/57800; acc:  68.40; ppl:  3.72; xent: 1.31; lr: 0.00038; 6831/6663 tok/s;  26784 sec\n",
            "[2020-10-24 01:24:57,741 INFO] Step 53150/57800; acc:  68.18; ppl:  3.74; xent: 1.32; lr: 0.00038; 6616/6716 tok/s;  26838 sec\n",
            "[2020-10-24 01:25:52,529 INFO] Step 53200/57800; acc:  68.67; ppl:  3.65; xent: 1.29; lr: 0.00038; 6550/6763 tok/s;  26893 sec\n",
            "[2020-10-24 01:26:47,599 INFO] Step 53250/57800; acc:  68.65; ppl:  3.64; xent: 1.29; lr: 0.00038; 6506/6694 tok/s;  26948 sec\n",
            "[2020-10-24 01:27:41,673 INFO] Step 53300/57800; acc:  69.00; ppl:  3.58; xent: 1.28; lr: 0.00038; 6588/6483 tok/s;  27002 sec\n",
            "[2020-10-24 01:28:36,221 INFO] Step 53350/57800; acc:  69.65; ppl:  3.48; xent: 1.25; lr: 0.00038; 6547/6782 tok/s;  27057 sec\n",
            "[2020-10-24 01:29:30,746 INFO] Step 53400/57800; acc:  69.08; ppl:  3.54; xent: 1.26; lr: 0.00038; 6721/6630 tok/s;  27111 sec\n",
            "[2020-10-24 01:30:25,149 INFO] Step 53450/57800; acc:  69.47; ppl:  3.50; xent: 1.25; lr: 0.00038; 6683/6686 tok/s;  27165 sec\n",
            "[2020-10-24 01:31:19,969 INFO] Step 53500/57800; acc:  69.34; ppl:  3.51; xent: 1.25; lr: 0.00038; 6690/6705 tok/s;  27220 sec\n",
            "[2020-10-24 01:32:12,750 INFO] Step 53550/57800; acc:  69.50; ppl:  3.47; xent: 1.24; lr: 0.00038; 6918/6644 tok/s;  27273 sec\n",
            "[2020-10-24 01:33:06,664 INFO] Step 53600/57800; acc:  69.30; ppl:  3.51; xent: 1.25; lr: 0.00038; 6748/6639 tok/s;  27327 sec\n",
            "[2020-10-24 01:34:01,156 INFO] Step 53650/57800; acc:  69.57; ppl:  3.45; xent: 1.24; lr: 0.00038; 6607/6696 tok/s;  27381 sec\n",
            "[2020-10-24 01:34:55,036 INFO] Step 53700/57800; acc:  69.75; ppl:  3.41; xent: 1.23; lr: 0.00038; 6775/6685 tok/s;  27435 sec\n",
            "[2020-10-24 01:35:49,283 INFO] Step 53750/57800; acc:  70.13; ppl:  3.35; xent: 1.21; lr: 0.00038; 6599/6692 tok/s;  27490 sec\n",
            "[2020-10-24 01:36:43,036 INFO] Step 53800/57800; acc:  70.21; ppl:  3.34; xent: 1.21; lr: 0.00038; 6773/6725 tok/s;  27543 sec\n",
            "[2020-10-24 01:37:36,777 INFO] Step 53850/57800; acc:  70.42; ppl:  3.31; xent: 1.20; lr: 0.00038; 6765/6651 tok/s;  27597 sec\n",
            "[2020-10-24 01:38:31,058 INFO] Step 53900/57800; acc:  70.62; ppl:  3.27; xent: 1.18; lr: 0.00038; 6665/6634 tok/s;  27651 sec\n",
            "[2020-10-24 01:39:24,620 INFO] Step 53950/57800; acc:  70.31; ppl:  3.31; xent: 1.20; lr: 0.00038; 6672/6650 tok/s;  27705 sec\n",
            "[2020-10-24 01:40:18,240 INFO] Step 54000/57800; acc:  71.26; ppl:  3.16; xent: 1.15; lr: 0.00038; 6589/6772 tok/s;  27759 sec\n",
            "[2020-10-24 01:41:12,633 INFO] Step 54050/57800; acc:  70.65; ppl:  3.24; xent: 1.18; lr: 0.00038; 6672/6659 tok/s;  27813 sec\n",
            "[2020-10-24 01:42:06,364 INFO] Step 54100/57800; acc:  70.33; ppl:  3.28; xent: 1.19; lr: 0.00038; 6715/6664 tok/s;  27867 sec\n",
            "[2020-10-24 01:42:59,631 INFO] Step 54150/57800; acc:  70.70; ppl:  3.23; xent: 1.17; lr: 0.00038; 6802/6656 tok/s;  27920 sec\n",
            "[2020-10-24 01:43:53,770 INFO] Step 54200/57800; acc:  70.69; ppl:  3.23; xent: 1.17; lr: 0.00038; 6724/6681 tok/s;  27974 sec\n",
            "[2020-10-24 01:44:47,368 INFO] Step 54250/57800; acc:  71.60; ppl:  3.09; xent: 1.13; lr: 0.00038; 6680/6589 tok/s;  28028 sec\n",
            "[2020-10-24 01:45:42,372 INFO] Step 54300/57800; acc:  71.05; ppl:  3.18; xent: 1.16; lr: 0.00038; 6518/6609 tok/s;  28083 sec\n",
            "[2020-10-24 01:46:36,919 INFO] Step 54350/57800; acc:  71.11; ppl:  3.16; xent: 1.15; lr: 0.00038; 6584/6692 tok/s;  28137 sec\n",
            "[2020-10-24 01:47:31,600 INFO] Step 54400/57800; acc:  71.19; ppl:  3.13; xent: 1.14; lr: 0.00038; 6513/6716 tok/s;  28192 sec\n",
            "[2020-10-24 01:48:26,708 INFO] Step 54450/57800; acc:  71.52; ppl:  3.10; xent: 1.13; lr: 0.00038; 6552/6754 tok/s;  28247 sec\n",
            "[2020-10-24 01:49:17,079 INFO] Loading dataset from data/multi-europarl.train.17.pt\n",
            "[2020-10-24 01:49:39,003 INFO] number of examples: 425752\n",
            "[2020-10-24 01:49:48,116 INFO] Step 54500/57800; acc:  71.11; ppl:  3.17; xent: 1.15; lr: 0.00038; 4488/4469 tok/s;  28328 sec\n",
            "[2020-10-24 01:50:41,893 INFO] Step 54550/57800; acc:  67.79; ppl:  3.82; xent: 1.34; lr: 0.00038; 6718/6569 tok/s;  28382 sec\n",
            "[2020-10-24 01:51:35,994 INFO] Step 54600/57800; acc:  68.43; ppl:  3.74; xent: 1.32; lr: 0.00038; 6642/6644 tok/s;  28436 sec\n",
            "[2020-10-24 01:52:29,716 INFO] Step 54650/57800; acc:  68.40; ppl:  3.70; xent: 1.31; lr: 0.00038; 6678/6652 tok/s;  28490 sec\n",
            "[2020-10-24 01:53:24,355 INFO] Step 54700/57800; acc:  68.52; ppl:  3.68; xent: 1.30; lr: 0.00038; 6542/6716 tok/s;  28545 sec\n",
            "[2020-10-24 01:54:19,705 INFO] Step 54750/57800; acc:  68.60; ppl:  3.64; xent: 1.29; lr: 0.00038; 6425/6717 tok/s;  28600 sec\n",
            "[2020-10-24 01:55:13,765 INFO] Step 54800/57800; acc:  68.83; ppl:  3.60; xent: 1.28; lr: 0.00038; 6695/6702 tok/s;  28654 sec\n",
            "[2020-10-24 01:56:07,873 INFO] Step 54850/57800; acc:  69.17; ppl:  3.54; xent: 1.26; lr: 0.00038; 6652/6664 tok/s;  28708 sec\n",
            "[2020-10-24 01:57:01,588 INFO] Step 54900/57800; acc:  69.61; ppl:  3.47; xent: 1.24; lr: 0.00038; 6707/6678 tok/s;  28762 sec\n",
            "[2020-10-24 01:57:56,129 INFO] Step 54950/57800; acc:  69.33; ppl:  3.50; xent: 1.25; lr: 0.00038; 6731/6685 tok/s;  28816 sec\n",
            "[2020-10-24 01:58:50,356 INFO] Step 55000/57800; acc:  69.50; ppl:  3.47; xent: 1.24; lr: 0.00038; 6611/6726 tok/s;  28871 sec\n",
            "[2020-10-24 01:59:44,782 INFO] Step 55050/57800; acc:  69.64; ppl:  3.44; xent: 1.24; lr: 0.00038; 6579/6628 tok/s;  28925 sec\n",
            "[2020-10-24 02:00:38,995 INFO] Step 55100/57800; acc:  69.79; ppl:  3.43; xent: 1.23; lr: 0.00038; 6830/6655 tok/s;  28979 sec\n",
            "[2020-10-24 02:01:33,707 INFO] Step 55150/57800; acc:  69.54; ppl:  3.45; xent: 1.24; lr: 0.00038; 6613/6625 tok/s;  29034 sec\n",
            "[2020-10-24 02:02:28,596 INFO] Step 55200/57800; acc:  69.87; ppl:  3.41; xent: 1.23; lr: 0.00038; 6625/6700 tok/s;  29089 sec\n",
            "[2020-10-24 02:03:22,558 INFO] Step 55250/57800; acc:  69.77; ppl:  3.41; xent: 1.23; lr: 0.00038; 6693/6705 tok/s;  29143 sec\n",
            "[2020-10-24 02:04:16,888 INFO] Step 55300/57800; acc:  69.99; ppl:  3.38; xent: 1.22; lr: 0.00038; 6614/6709 tok/s;  29197 sec\n",
            "[2020-10-24 02:05:11,702 INFO] Step 55350/57800; acc:  70.44; ppl:  3.31; xent: 1.20; lr: 0.00038; 6687/6720 tok/s;  29252 sec\n",
            "[2020-10-24 02:06:05,114 INFO] Step 55400/57800; acc:  70.29; ppl:  3.34; xent: 1.21; lr: 0.00038; 6781/6618 tok/s;  29305 sec\n",
            "[2020-10-24 02:06:59,950 INFO] Step 55450/57800; acc:  70.76; ppl:  3.26; xent: 1.18; lr: 0.00038; 6568/6694 tok/s;  29360 sec\n",
            "[2020-10-24 02:07:52,904 INFO] Step 55500/57800; acc:  70.49; ppl:  3.28; xent: 1.19; lr: 0.00038; 6864/6566 tok/s;  29413 sec\n",
            "[2020-10-24 02:08:47,453 INFO] Step 55550/57800; acc:  71.01; ppl:  3.20; xent: 1.16; lr: 0.00038; 6497/6602 tok/s;  29468 sec\n",
            "[2020-10-24 02:09:41,618 INFO] Step 55600/57800; acc:  70.98; ppl:  3.20; xent: 1.16; lr: 0.00037; 6675/6723 tok/s;  29522 sec\n",
            "[2020-10-24 02:10:35,529 INFO] Step 55650/57800; acc:  70.40; ppl:  3.27; xent: 1.18; lr: 0.00037; 6758/6688 tok/s;  29576 sec\n",
            "[2020-10-24 02:11:28,898 INFO] Step 55700/57800; acc:  70.56; ppl:  3.23; xent: 1.17; lr: 0.00037; 6768/6659 tok/s;  29629 sec\n",
            "[2020-10-24 02:12:23,041 INFO] Step 55750/57800; acc:  70.84; ppl:  3.21; xent: 1.17; lr: 0.00037; 6686/6679 tok/s;  29683 sec\n",
            "[2020-10-24 02:13:17,450 INFO] Step 55800/57800; acc:  71.01; ppl:  3.17; xent: 1.15; lr: 0.00037; 6647/6638 tok/s;  29738 sec\n",
            "[2020-10-24 02:14:11,814 INFO] Step 55850/57800; acc:  71.00; ppl:  3.18; xent: 1.16; lr: 0.00037; 6618/6715 tok/s;  29792 sec\n",
            "[2020-10-24 02:15:06,314 INFO] Step 55900/57800; acc:  71.36; ppl:  3.12; xent: 1.14; lr: 0.00037; 6633/6678 tok/s;  29847 sec\n",
            "[2020-10-24 02:16:00,442 INFO] Step 55950/57800; acc:  71.57; ppl:  3.10; xent: 1.13; lr: 0.00037; 6713/6679 tok/s;  29901 sec\n",
            "[2020-10-24 02:16:54,329 INFO] Step 56000/57800; acc:  71.46; ppl:  3.10; xent: 1.13; lr: 0.00037; 6594/6685 tok/s;  29955 sec\n",
            "[2020-10-24 02:16:54,394 INFO] Loading dataset from data/multi-europarl.train.18.pt\n",
            "[2020-10-24 02:17:16,727 INFO] number of examples: 426339\n",
            "[2020-10-24 02:18:16,675 INFO] Step 56050/57800; acc:  67.91; ppl:  3.79; xent: 1.33; lr: 0.00037; 4413/4528 tok/s;  30037 sec\n",
            "[2020-10-24 02:19:11,604 INFO] Step 56100/57800; acc:  68.29; ppl:  3.72; xent: 1.31; lr: 0.00037; 6612/6678 tok/s;  30092 sec\n",
            "[2020-10-24 02:20:05,236 INFO] Step 56150/57800; acc:  68.54; ppl:  3.68; xent: 1.30; lr: 0.00037; 6778/6653 tok/s;  30146 sec\n",
            "[2020-10-24 02:20:59,131 INFO] Step 56200/57800; acc:  68.65; ppl:  3.65; xent: 1.30; lr: 0.00037; 6567/6706 tok/s;  30199 sec\n",
            "[2020-10-24 02:21:53,277 INFO] Step 56250/57800; acc:  69.31; ppl:  3.53; xent: 1.26; lr: 0.00037; 6685/6729 tok/s;  30254 sec\n",
            "[2020-10-24 02:22:47,562 INFO] Step 56300/57800; acc:  69.10; ppl:  3.56; xent: 1.27; lr: 0.00037; 6682/6726 tok/s;  30308 sec\n",
            "[2020-10-24 02:23:41,999 INFO] Step 56350/57800; acc:  69.08; ppl:  3.55; xent: 1.27; lr: 0.00037; 6726/6626 tok/s;  30362 sec\n",
            "[2020-10-24 02:24:35,403 INFO] Step 56400/57800; acc:  69.16; ppl:  3.53; xent: 1.26; lr: 0.00037; 6736/6563 tok/s;  30416 sec\n",
            "[2020-10-24 02:25:30,038 INFO] Step 56450/57800; acc:  69.46; ppl:  3.48; xent: 1.25; lr: 0.00037; 6665/6675 tok/s;  30470 sec\n",
            "[2020-10-24 02:26:23,522 INFO] Step 56500/57800; acc:  69.48; ppl:  3.47; xent: 1.24; lr: 0.00037; 6734/6699 tok/s;  30524 sec\n",
            "[2020-10-24 02:27:17,839 INFO] Step 56550/57800; acc:  69.26; ppl:  3.49; xent: 1.25; lr: 0.00037; 6667/6648 tok/s;  30578 sec\n",
            "[2020-10-24 02:28:12,100 INFO] Step 56600/57800; acc:  69.82; ppl:  3.41; xent: 1.23; lr: 0.00037; 6642/6671 tok/s;  30632 sec\n",
            "[2020-10-24 02:29:05,529 INFO] Step 56650/57800; acc:  70.09; ppl:  3.37; xent: 1.21; lr: 0.00037; 6894/6650 tok/s;  30686 sec\n",
            "[2020-10-24 02:29:59,353 INFO] Step 56700/57800; acc:  70.16; ppl:  3.36; xent: 1.21; lr: 0.00037; 6753/6722 tok/s;  30740 sec\n",
            "[2020-10-24 02:30:53,342 INFO] Step 56750/57800; acc:  69.81; ppl:  3.40; xent: 1.22; lr: 0.00037; 6610/6574 tok/s;  30794 sec\n",
            "[2020-10-24 02:31:47,351 INFO] Step 56800/57800; acc:  70.14; ppl:  3.33; xent: 1.20; lr: 0.00037; 6711/6697 tok/s;  30848 sec\n",
            "[2020-10-24 02:32:41,040 INFO] Step 56850/57800; acc:  70.42; ppl:  3.30; xent: 1.19; lr: 0.00037; 6688/6649 tok/s;  30901 sec\n",
            "[2020-10-24 02:33:34,476 INFO] Step 56900/57800; acc:  70.83; ppl:  3.23; xent: 1.17; lr: 0.00037; 6698/6651 tok/s;  30955 sec\n",
            "[2020-10-24 02:34:28,825 INFO] Step 56950/57800; acc:  70.32; ppl:  3.32; xent: 1.20; lr: 0.00037; 6460/6634 tok/s;  31009 sec\n",
            "[2020-10-24 02:35:23,526 INFO] Step 57000/57800; acc:  70.81; ppl:  3.23; xent: 1.17; lr: 0.00037; 6641/6701 tok/s;  31064 sec\n",
            "[2020-10-24 02:36:19,223 INFO] Step 57050/57800; acc:  70.24; ppl:  3.31; xent: 1.20; lr: 0.00037; 6498/6657 tok/s;  31120 sec\n",
            "[2020-10-24 02:37:13,432 INFO] Step 57100/57800; acc:  70.87; ppl:  3.20; xent: 1.16; lr: 0.00037; 6532/6717 tok/s;  31174 sec\n",
            "[2020-10-24 02:38:07,276 INFO] Step 57150/57800; acc:  70.92; ppl:  3.20; xent: 1.16; lr: 0.00037; 6635/6739 tok/s;  31228 sec\n",
            "[2020-10-24 02:39:01,318 INFO] Step 57200/57800; acc:  71.06; ppl:  3.17; xent: 1.15; lr: 0.00037; 6720/6761 tok/s;  31282 sec\n",
            "[2020-10-24 02:39:55,246 INFO] Step 57250/57800; acc:  70.58; ppl:  3.25; xent: 1.18; lr: 0.00037; 6817/6658 tok/s;  31336 sec\n",
            "[2020-10-24 02:40:48,475 INFO] Step 57300/57800; acc:  71.26; ppl:  3.14; xent: 1.15; lr: 0.00037; 6831/6584 tok/s;  31389 sec\n",
            "[2020-10-24 02:41:42,747 INFO] Step 57350/57800; acc:  71.28; ppl:  3.13; xent: 1.14; lr: 0.00037; 6661/6767 tok/s;  31443 sec\n",
            "[2020-10-24 02:42:37,257 INFO] Step 57400/57800; acc:  71.82; ppl:  3.08; xent: 1.12; lr: 0.00037; 6628/6658 tok/s;  31498 sec\n",
            "[2020-10-24 02:43:31,227 INFO] Step 57450/57800; acc:  71.47; ppl:  3.09; xent: 1.13; lr: 0.00037; 6609/6686 tok/s;  31552 sec\n",
            "[2020-10-24 02:44:26,008 INFO] Step 57500/57800; acc:  70.91; ppl:  3.16; xent: 1.15; lr: 0.00037; 6516/6626 tok/s;  31606 sec\n",
            "[2020-10-24 02:44:28,267 INFO] Loading dataset from data/multi-europarl.train.19.pt\n",
            "[2020-10-24 02:44:32,140 INFO] number of examples: 85557\n",
            "[2020-10-24 02:45:26,606 INFO] Step 57550/57800; acc:  68.32; ppl:  3.72; xent: 1.31; lr: 0.00037; 5950/5835 tok/s;  31667 sec\n",
            "[2020-10-24 02:46:22,208 INFO] Step 57600/57800; acc:  69.40; ppl:  3.52; xent: 1.26; lr: 0.00037; 6242/6456 tok/s;  31723 sec\n",
            "[2020-10-24 02:47:17,333 INFO] Step 57650/57800; acc:  70.11; ppl:  3.39; xent: 1.22; lr: 0.00037; 6310/6377 tok/s;  31778 sec\n",
            "[2020-10-24 02:48:12,120 INFO] Step 57700/57800; acc:  70.62; ppl:  3.28; xent: 1.19; lr: 0.00037; 6525/6361 tok/s;  31832 sec\n",
            "[2020-10-24 02:49:06,633 INFO] Step 57750/57800; acc:  71.38; ppl:  3.15; xent: 1.15; lr: 0.00037; 6525/6458 tok/s;  31887 sec\n",
            "[2020-10-24 02:50:01,528 INFO] Step 57800/57800; acc:  71.64; ppl:  3.09; xent: 1.13; lr: 0.00037; 6250/6352 tok/s;  31942 sec\n",
            "[2020-10-24 02:50:02,249 INFO] Saving checkpoint models/multi-europarl-v3_step_57800.pt\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9Dj_8tpXLhcP"
      },
      "source": [
        "# Translating"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XtU8_0CGLthv",
        "outputId": "2dfb076f-70e6-4caf-c767-ab4701adfd55",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "!onmt_translate -model models/multi-europarl_step_200000_dbpe_v2.pt -src data/europarl_test_src.txt -output dbpe_pred_v2_200k.txt -gpu 0"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[2020-10-07 18:22:58,592 INFO] Translating shard 0.\n",
            "[2020-10-07 18:27:32,020 INFO] PRED AVG SCORE: -0.4348, PRED PPL: 1.5446\n",
            "[2020-10-07 18:27:32,058 INFO] Translating shard 1.\n",
            "[2020-10-07 18:32:18,199 INFO] PRED AVG SCORE: -0.4375, PRED PPL: 1.5488\n",
            "[2020-10-07 18:32:18,237 INFO] Translating shard 2.\n",
            "[2020-10-07 18:36:59,945 INFO] PRED AVG SCORE: -0.4293, PRED PPL: 1.5361\n",
            "[2020-10-07 18:36:59,983 INFO] Translating shard 3.\n",
            "[2020-10-07 18:41:33,540 INFO] PRED AVG SCORE: -0.4323, PRED PPL: 1.5408\n",
            "[2020-10-07 18:41:33,575 INFO] Translating shard 4.\n",
            "[2020-10-07 18:46:17,955 INFO] PRED AVG SCORE: -0.4324, PRED PPL: 1.5409\n",
            "[2020-10-07 18:46:17,990 INFO] Translating shard 5.\n",
            "[2020-10-07 18:50:48,252 INFO] PRED AVG SCORE: -0.4304, PRED PPL: 1.5378\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IS5QahD8lXw6"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}