training_grnn.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4",
      "collapsed_sections": [
        "O2iahSd2N0WY",
        "x7XN2DeWN4oq",
        "O5QcmUEbN7si",
        "sgLBW26sN9rY",
        "taeH4HZlOEWm"
      ]
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "## Clone the repo"
      ],
      "metadata": {
        "id": "O2iahSd2N0WY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!rm -r ./*\n",
        "!git clone https://github.com/Mamiglia/BNN_Human_motion\n",
        "!mv BNN_Human_motion/* .\n",
        "!pip install bayesian-torch lightning"
      ],
      "metadata": {
        "id": "kQytzkH_0ZBP"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Needed import"
      ],
      "metadata": {
        "id": "x7XN2DeWN4oq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import torch\n",
        "import os\n",
        "from funcs.dataloader import load_dataset\n",
        "from funcs.gated_recurrent_nn.model import Model\n",
        "from funcs.gated_recurrent_nn.training import Training\n",
        "from funcs.gated_recurrent_nn.testing import *"
      ],
      "metadata": {
        "id": "YjQeUc0VNtcW"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Set the device"
      ],
      "metadata": {
        "id": "O5QcmUEbN7si"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
        "print('Using device:', device, '- Type:', torch.cuda.get_device_name(0))"
      ],
      "metadata": {
        "id": "Xeptqw0gNioz",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "29cac7df-ecd2-432e-dc39-035d11db8870"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Using device: cuda - Type: Tesla T4\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Load the datasets"
      ],
      "metadata": {
        "id": "sgLBW26sN9rY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_dataset, vald_dataset, _ = load_dataset()"
      ],
      "metadata": {
        "id": "m1tcGj8ZNngF",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "73a43267-45ae-4f52-8a9d-d904872d0704"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Loading Train Dataset...\n",
            "Loading Validation Dataset...\n",
            "Loading Test Dataset...\n",
            ">>> Training dataset length: 180077\n",
            ">>> Validation dataset length: 28110\n",
            ">>> Test dataset length: 3840\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model initialization"
      ],
      "metadata": {
        "id": "taeH4HZlOEWm"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "num_channels = 3\n",
        "output_n = 25\n",
        "input_n = 10\n",
        "num_joints = 22\n",
        "num_heads = 8\n",
        "dropout = 0.3\n",
        "d_model = 512\n",
        "num_predictions = 2\n",
        "\n",
        "model = Model(num_channels=num_channels,\n",
        "              num_frames_out=output_n,\n",
        "              old_frames=input_n,\n",
        "              num_joints=num_joints,\n",
        "              num_heads=num_heads,\n",
        "              drop=dropout,\n",
        "              d_model=d_model,\n",
        "              num_predictions=num_predictions).to(device)"
      ],
      "metadata": {
        "id": "ACNncjkZNzNK"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "trainable, non_trainable, total = model.count_parameters()\n",
        "print(f'Trainable parameters: {trainable}')\n",
        "print(f'Total parameters: {total}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XVt5nGDfzWwU",
        "outputId": "108656b0-9d1c-4fa7-8c59-af6ee3b251bd"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Trainable parameters: 5281177\n",
            "Total parameters: 5281177\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Training"
      ],
      "metadata": {
        "id": "YjHRtSe0O6Vg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install wandb -qqq\n",
        "import wandb\n",
        "wandb.login()"
      ],
      "metadata": {
        "id": "lOBeOnJeKXrq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "clip_grad = None\n",
        "n_epochs = 2\n",
        "log_step = 100\n",
        "lr = 1e-04\n",
        "use_scheduler = True\n",
        "milestones = [4, 8, 12, 16]\n",
        "gamma = 0.7\n",
        "weight_decay = 3e-04\n",
        "use_wandb = False\n",
        "save_and_plot = True\n",
        "path = os.path.join(os.getcwd(), 'funcs/gated_recurrent_nn/checkpoints/Best_checkpoint.pt')\n",
        "\n",
        "\n",
        "if use_wandb:\n",
        "    wandb.init(project=\"AML_FinalProject\", config={\n",
        "        \"learning_rate\": lr,\n",
        "        \"dropout\": dropout,\n",
        "        \"gamma\": gamma,\n",
        "        \"weight_decay\": weight_decay,\n",
        "        \"milestones\": milestones,\n",
        "        \"architecture\": \"Encoder (Transformer+CNN) Decoder (GRU+CNN)\",\n",
        "        \"parameters\": trainable,\n",
        "    })\n",
        "    config = wandb.config\n",
        "\n",
        "\n",
        "model_training = Training(model=model,\n",
        "                          input_n=input_n,\n",
        "                          output_n=output_n,\n",
        "                          clip_grad=clip_grad,\n",
        "                          device=device,\n",
        "                          n_epochs=n_epochs,\n",
        "                          log_step=log_step,\n",
        "                          lr=lr,\n",
        "                          use_scheduler=use_scheduler,\n",
        "                          milestones=milestones,\n",
        "                          gamma=gamma,\n",
        "                          weight_decay=weight_decay,\n",
        "                          use_wandb=use_wandb,\n",
        "                          save_and_plot=save_and_plot,\n",
        "                          path=path)"
      ],
      "metadata": {
        "id": "tseD_HzvNt2n"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "gn1COwQEMUwO",
        "outputId": "ee458296-804c-42e8-ed76-8d8d31e4795a"
      },
      "outputs": [
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "\r0it [00:00, ?it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 1  | Iteration:     1 | Train ] MPJPE loss: 107.290 | KL loss: 0.989 | Total loss: 108.279\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100it [02:30,  1.50s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 1  | Iteration:   101 | Train ] MPJPE loss: 99.251 | KL loss: 0.985 | Total loss: 100.236\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "200it [05:00,  1.55s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 1  | Iteration:   201 | Train ] MPJPE loss: 105.303 | KL loss: 0.981 | Total loss: 106.285\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "300it [07:30,  1.48s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 1  | Iteration:   301 | Train ] MPJPE loss: 102.230 | KL loss: 0.978 | Total loss: 103.207\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "400it [09:59,  1.52s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 1  | Iteration:   401 | Train ] MPJPE loss: 99.491 | KL loss: 0.974 | Total loss: 100.465\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "500it [12:29,  1.50s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 1  | Iteration:   501 | Train ] MPJPE loss: 105.486 | KL loss: 0.970 | Total loss: 106.456\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "600it [14:59,  1.49s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 1  | Iteration:   601 | Train ] MPJPE loss: 104.635 | KL loss: 0.967 | Total loss: 105.601\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "700it [17:29,  1.51s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 1  | Iteration:   701 | Train ] MPJPE loss: 95.562 | KL loss: 0.963 | Total loss: 96.525\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "704it [17:35,  1.50s/it]\n",
            "1it [00:01,  1.06s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m[Epoch: 1  | Iteration:     1 | Val   ] MPJPE loss: 118.894 | KL loss: 0.963 | Total loss: 119.857\u001b[0m\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "101it [00:48,  2.13it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m[Epoch: 1  | Iteration:   101 | Val   ] MPJPE loss: 93.423 | KL loss: 0.963 | Total loss: 94.386\u001b[0m\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "110it [00:53,  2.05it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 2  | Iteration:     1 | Train ] MPJPE loss: 99.391 | KL loss: 0.963 | Total loss: 100.354\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100it [02:36,  1.55s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 2  | Iteration:   101 | Train ] MPJPE loss: 97.228 | KL loss: 0.960 | Total loss: 98.188\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "200it [05:12,  1.55s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 2  | Iteration:   201 | Train ] MPJPE loss: 90.856 | KL loss: 0.957 | Total loss: 91.813\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "300it [07:46,  1.52s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 2  | Iteration:   301 | Train ] MPJPE loss: 97.300 | KL loss: 0.954 | Total loss: 98.253\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "400it [10:19,  1.51s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 2  | Iteration:   401 | Train ] MPJPE loss: 90.447 | KL loss: 0.951 | Total loss: 91.398\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "500it [12:52,  1.53s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 2  | Iteration:   501 | Train ] MPJPE loss: 96.305 | KL loss: 0.948 | Total loss: 97.253\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "600it [15:25,  1.53s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 2  | Iteration:   601 | Train ] MPJPE loss: 92.054 | KL loss: 0.945 | Total loss: 92.999\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "700it [17:59,  1.55s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 2  | Iteration:   701 | Train ] MPJPE loss: 95.530 | KL loss: 0.942 | Total loss: 96.472\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "704it [18:05,  1.54s/it]\n",
            "1it [00:00,  1.21it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m[Epoch: 2  | Iteration:     1 | Val   ] MPJPE loss: 107.413 | KL loss: 0.942 | Total loss: 108.354\u001b[0m\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "101it [00:48,  2.12it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m[Epoch: 2  | Iteration:   101 | Val   ] MPJPE loss: 95.791 | KL loss: 0.942 | Total loss: 96.733\u001b[0m\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "110it [00:52,  2.08it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 3  | Iteration:     1 | Train ] MPJPE loss: 93.483 | KL loss: 0.942 | Total loss: 94.425\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100it [02:38,  1.60s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 3  | Iteration:   101 | Train ] MPJPE loss: 87.465 | KL loss: 0.939 | Total loss: 88.404\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "200it [05:16,  1.59s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 3  | Iteration:   201 | Train ] MPJPE loss: 92.503 | KL loss: 0.936 | Total loss: 93.439\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "300it [07:52,  1.56s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 3  | Iteration:   301 | Train ] MPJPE loss: 98.028 | KL loss: 0.933 | Total loss: 98.961\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "400it [10:27,  1.54s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 3  | Iteration:   401 | Train ] MPJPE loss: 90.373 | KL loss: 0.930 | Total loss: 91.303\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "500it [13:01,  1.54s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 3  | Iteration:   501 | Train ] MPJPE loss: 100.479 | KL loss: 0.927 | Total loss: 101.406\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "600it [15:34,  1.53s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 3  | Iteration:   601 | Train ] MPJPE loss: 88.639 | KL loss: 0.925 | Total loss: 89.563\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "700it [18:07,  1.54s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 3  | Iteration:   701 | Train ] MPJPE loss: 94.344 | KL loss: 0.922 | Total loss: 95.266\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "704it [18:14,  1.55s/it]\n",
            "1it [00:00,  1.23it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m[Epoch: 3  | Iteration:     1 | Val   ] MPJPE loss: 106.476 | KL loss: 0.922 | Total loss: 107.398\u001b[0m\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "101it [00:48,  2.12it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m[Epoch: 3  | Iteration:   101 | Val   ] MPJPE loss: 93.892 | KL loss: 0.922 | Total loss: 94.814\u001b[0m\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "110it [00:53,  2.04it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 4  | Iteration:     1 | Train ] MPJPE loss: 98.691 | KL loss: 0.922 | Total loss: 99.612\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100it [02:39,  1.59s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 4  | Iteration:   101 | Train ] MPJPE loss: 90.255 | KL loss: 0.919 | Total loss: 91.174\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "200it [05:17,  1.57s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 4  | Iteration:   201 | Train ] MPJPE loss: 97.017 | KL loss: 0.916 | Total loss: 97.933\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "300it [07:53,  1.55s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 4  | Iteration:   301 | Train ] MPJPE loss: 95.521 | KL loss: 0.913 | Total loss: 96.434\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "400it [10:28,  1.55s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 4  | Iteration:   401 | Train ] MPJPE loss: 92.112 | KL loss: 0.910 | Total loss: 93.022\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "500it [13:02,  1.54s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 4  | Iteration:   501 | Train ] MPJPE loss: 93.325 | KL loss: 0.908 | Total loss: 94.232\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "600it [15:37,  1.61s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 4  | Iteration:   601 | Train ] MPJPE loss: 91.801 | KL loss: 0.905 | Total loss: 92.705\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "700it [18:10,  1.56s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 4  | Iteration:   701 | Train ] MPJPE loss: 94.279 | KL loss: 0.902 | Total loss: 95.181\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "704it [18:17,  1.56s/it]\n",
            "1it [00:00,  1.15it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m[Epoch: 4  | Iteration:     1 | Val   ] MPJPE loss: 103.428 | KL loss: 0.902 | Total loss: 104.330\u001b[0m\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "101it [00:48,  2.08it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m[Epoch: 4  | Iteration:   101 | Val   ] MPJPE loss: 94.011 | KL loss: 0.902 | Total loss: 94.913\u001b[0m\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "110it [00:53,  2.04it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 5  | Iteration:     1 | Train ] MPJPE loss: 90.107 | KL loss: 0.902 | Total loss: 91.009\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100it [02:38,  1.59s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 5  | Iteration:   101 | Train ] MPJPE loss: 99.285 | KL loss: 0.899 | Total loss: 100.184\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "200it [05:16,  1.56s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 5  | Iteration:   201 | Train ] MPJPE loss: 94.395 | KL loss: 0.896 | Total loss: 95.292\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "300it [07:51,  1.53s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 5  | Iteration:   301 | Train ] MPJPE loss: 91.859 | KL loss: 0.893 | Total loss: 92.752\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "400it [10:26,  1.54s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 5  | Iteration:   401 | Train ] MPJPE loss: 90.272 | KL loss: 0.891 | Total loss: 91.163\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "500it [12:59,  1.52s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 5  | Iteration:   501 | Train ] MPJPE loss: 88.511 | KL loss: 0.888 | Total loss: 89.399\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "600it [15:32,  1.53s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 5  | Iteration:   601 | Train ] MPJPE loss: 89.132 | KL loss: 0.885 | Total loss: 90.017\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "700it [18:05,  1.54s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 5  | Iteration:   701 | Train ] MPJPE loss: 91.259 | KL loss: 0.882 | Total loss: 92.141\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "704it [18:11,  1.55s/it]\n",
            "1it [00:00,  1.63it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m[Epoch: 5  | Iteration:     1 | Val   ] MPJPE loss: 99.078 | KL loss: 0.882 | Total loss: 99.960\u001b[0m\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "101it [00:47,  2.06it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m[Epoch: 5  | Iteration:   101 | Val   ] MPJPE loss: 93.684 | KL loss: 0.882 | Total loss: 94.567\u001b[0m\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "110it [00:52,  2.10it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 6  | Iteration:     1 | Train ] MPJPE loss: 91.637 | KL loss: 0.882 | Total loss: 92.520\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100it [02:37,  1.58s/it]"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[Epoch: 6  | Iteration:   101 | Train ] MPJPE loss: 88.720 | KL loss: 0.880 | Total loss: 89.599\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:15,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 6  | Iteration:   201 | Train ] MPJPE loss: 91.691 | KL loss: 0.877 | Total loss: 92.568\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:50,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 6  | Iteration:   301 | Train ] MPJPE loss: 87.276 | KL loss: 0.874 | Total loss: 88.150\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:24,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 6  | Iteration:   401 | Train ] MPJPE loss: 88.402 | KL loss: 0.871 | Total loss: 89.273\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [12:58,  1.58s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 6  | Iteration:   501 | Train ] MPJPE loss: 85.857 | KL loss: 0.869 | Total loss: 86.725\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:31,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 6  | Iteration:   601 | Train ] MPJPE loss: 93.057 | KL loss: 0.866 | Total loss: 93.923\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:04,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 6  | Iteration:   701 | Train ] MPJPE loss: 81.536 | KL loss: 0.863 | Total loss: 82.399\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:11,  1.55s/it]\n",
            "1it [00:00,  1.15it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 6  | Iteration:     1 | Val   ] MPJPE loss: 97.907 | KL loss: 0.863 | Total loss: 98.770\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:48,  2.11it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 6  | Iteration:   101 | Val   ] MPJPE loss: 92.916 | KL loss: 0.863 | Total loss: 93.779\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:53,  2.05it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 7  | Iteration:     1 | Train ] MPJPE loss: 89.250 | KL loss: 0.863 | Total loss: 90.113\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:38,  1.57s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 7  | Iteration:   101 | Train ] MPJPE loss: 93.513 | KL loss: 0.860 | Total loss: 94.374\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:16,  1.58s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 7  | Iteration:   201 | Train ] MPJPE loss: 92.562 | KL loss: 0.858 | Total loss: 93.420\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:51,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 7  | Iteration:   301 | Train ] MPJPE loss: 89.314 | KL loss: 0.855 | Total loss: 90.169\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:26,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 7  | Iteration:   401 | Train ] MPJPE loss: 85.862 | KL loss: 0.852 | Total loss: 86.714\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [13:01,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 7  | Iteration:   501 | Train ] MPJPE loss: 87.047 | KL loss: 0.850 | Total loss: 87.897\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:35,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 7  | Iteration:   601 | Train ] MPJPE loss: 88.552 | KL loss: 0.847 | Total loss: 89.398\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:08,  1.55s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 7  | Iteration:   701 | Train ] MPJPE loss: 90.490 | KL loss: 0.844 | Total loss: 91.335\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:15,  1.56s/it]\n",
            "1it [00:00,  1.19it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 7  | Iteration:     1 | Val   ] MPJPE loss: 94.965 | KL loss: 0.844 | Total loss: 95.809\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:49,  2.06it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 7  | Iteration:   101 | Val   ] MPJPE loss: 93.246 | KL loss: 0.844 | Total loss: 94.090\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:55,  1.99it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 8  | Iteration:     1 | Train ] MPJPE loss: 94.097 | KL loss: 0.844 | Total loss: 94.941\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:40,  1.59s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 8  | Iteration:   101 | Train ] MPJPE loss: 85.580 | KL loss: 0.841 | Total loss: 86.421\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:18,  1.58s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 8  | Iteration:   201 | Train ] MPJPE loss: 90.558 | KL loss: 0.839 | Total loss: 91.396\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:55,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 8  | Iteration:   301 | Train ] MPJPE loss: 90.414 | KL loss: 0.836 | Total loss: 91.249\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:31,  1.60s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 8  | Iteration:   401 | Train ] MPJPE loss: 82.578 | KL loss: 0.833 | Total loss: 83.411\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [13:06,  1.55s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 8  | Iteration:   501 | Train ] MPJPE loss: 94.255 | KL loss: 0.831 | Total loss: 95.086\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:40,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 8  | Iteration:   601 | Train ] MPJPE loss: 85.530 | KL loss: 0.828 | Total loss: 86.358\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:13,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 8  | Iteration:   701 | Train ] MPJPE loss: 86.636 | KL loss: 0.825 | Total loss: 87.461\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:20,  1.56s/it]\n",
            "1it [00:00,  1.14it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 8  | Iteration:     1 | Val   ] MPJPE loss: 91.164 | KL loss: 0.825 | Total loss: 91.989\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:48,  2.08it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 8  | Iteration:   101 | Val   ] MPJPE loss: 92.812 | KL loss: 0.825 | Total loss: 93.637\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:54,  2.03it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 9  | Iteration:     1 | Train ] MPJPE loss: 91.977 | KL loss: 0.825 | Total loss: 92.802\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:39,  1.58s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 9  | Iteration:   101 | Train ] MPJPE loss: 86.483 | KL loss: 0.822 | Total loss: 87.306\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:18,  1.57s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 9  | Iteration:   201 | Train ] MPJPE loss: 88.794 | KL loss: 0.820 | Total loss: 89.614\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:55,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 9  | Iteration:   301 | Train ] MPJPE loss: 92.130 | KL loss: 0.817 | Total loss: 92.947\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:30,  1.55s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 9  | Iteration:   401 | Train ] MPJPE loss: 84.479 | KL loss: 0.814 | Total loss: 85.293\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [13:04,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 9  | Iteration:   501 | Train ] MPJPE loss: 91.306 | KL loss: 0.812 | Total loss: 92.118\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:38,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 9  | Iteration:   601 | Train ] MPJPE loss: 84.996 | KL loss: 0.809 | Total loss: 85.805\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:11,  1.55s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 9  | Iteration:   701 | Train ] MPJPE loss: 85.760 | KL loss: 0.806 | Total loss: 86.567\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:18,  1.56s/it]\n",
            "1it [00:00,  1.13it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 9  | Iteration:     1 | Val   ] MPJPE loss: 89.615 | KL loss: 0.806 | Total loss: 90.421\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:49,  2.08it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 9  | Iteration:   101 | Val   ] MPJPE loss: 92.937 | KL loss: 0.806 | Total loss: 93.744\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:54,  2.02it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 10 | Iteration:     1 | Train ] MPJPE loss: 88.934 | KL loss: 0.806 | Total loss: 89.740\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:37,  1.59s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 10 | Iteration:   101 | Train ] MPJPE loss: 91.605 | KL loss: 0.803 | Total loss: 92.409\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:14,  1.59s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 10 | Iteration:   201 | Train ] MPJPE loss: 86.191 | KL loss: 0.801 | Total loss: 86.992\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:50,  1.57s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 10 | Iteration:   301 | Train ] MPJPE loss: 90.596 | KL loss: 0.798 | Total loss: 91.394\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:24,  1.55s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 10 | Iteration:   401 | Train ] MPJPE loss: 85.370 | KL loss: 0.795 | Total loss: 86.166\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [12:58,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 10 | Iteration:   501 | Train ] MPJPE loss: 90.168 | KL loss: 0.793 | Total loss: 90.961\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:32,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 10 | Iteration:   601 | Train ] MPJPE loss: 85.572 | KL loss: 0.790 | Total loss: 86.363\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:06,  1.55s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 10 | Iteration:   701 | Train ] MPJPE loss: 88.968 | KL loss: 0.787 | Total loss: 89.755\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:13,  1.55s/it]\n",
            "1it [00:00,  1.16it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 10 | Iteration:     1 | Val   ] MPJPE loss: 86.931 | KL loss: 0.787 | Total loss: 87.719\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:48,  2.10it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 10 | Iteration:   101 | Val   ] MPJPE loss: 92.474 | KL loss: 0.787 | Total loss: 93.262\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:54,  2.03it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 11 | Iteration:     1 | Train ] MPJPE loss: 92.343 | KL loss: 0.787 | Total loss: 93.131\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:39,  1.60s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 11 | Iteration:   101 | Train ] MPJPE loss: 90.764 | KL loss: 0.785 | Total loss: 91.549\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:18,  1.58s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 11 | Iteration:   201 | Train ] MPJPE loss: 89.349 | KL loss: 0.782 | Total loss: 90.132\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:56,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 11 | Iteration:   301 | Train ] MPJPE loss: 83.101 | KL loss: 0.779 | Total loss: 83.881\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:31,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 11 | Iteration:   401 | Train ] MPJPE loss: 85.503 | KL loss: 0.777 | Total loss: 86.279\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [13:05,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 11 | Iteration:   501 | Train ] MPJPE loss: 86.172 | KL loss: 0.774 | Total loss: 86.946\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:40,  1.59s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 11 | Iteration:   601 | Train ] MPJPE loss: 91.329 | KL loss: 0.771 | Total loss: 92.100\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:12,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 11 | Iteration:   701 | Train ] MPJPE loss: 87.632 | KL loss: 0.769 | Total loss: 88.401\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:19,  1.56s/it]\n",
            "1it [00:00,  1.18it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 11 | Iteration:     1 | Val   ] MPJPE loss: 85.668 | KL loss: 0.769 | Total loss: 86.436\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:48,  2.09it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 11 | Iteration:   101 | Val   ] MPJPE loss: 91.552 | KL loss: 0.769 | Total loss: 92.321\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:53,  2.05it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 12 | Iteration:     1 | Train ] MPJPE loss: 83.736 | KL loss: 0.769 | Total loss: 84.505\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:39,  1.58s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 12 | Iteration:   101 | Train ] MPJPE loss: 87.711 | KL loss: 0.766 | Total loss: 88.477\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:17,  1.59s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 12 | Iteration:   201 | Train ] MPJPE loss: 82.508 | KL loss: 0.763 | Total loss: 83.271\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:54,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 12 | Iteration:   301 | Train ] MPJPE loss: 86.635 | KL loss: 0.761 | Total loss: 87.396\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:29,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 12 | Iteration:   401 | Train ] MPJPE loss: 85.063 | KL loss: 0.758 | Total loss: 85.821\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [13:03,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 12 | Iteration:   501 | Train ] MPJPE loss: 86.132 | KL loss: 0.756 | Total loss: 86.888\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:36,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 12 | Iteration:   601 | Train ] MPJPE loss: 89.563 | KL loss: 0.753 | Total loss: 90.316\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:10,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 12 | Iteration:   701 | Train ] MPJPE loss: 85.413 | KL loss: 0.750 | Total loss: 86.163\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:16,  1.56s/it]\n",
            "1it [00:00,  1.16it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 12 | Iteration:     1 | Val   ] MPJPE loss: 85.078 | KL loss: 0.750 | Total loss: 85.828\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:48,  2.10it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 12 | Iteration:   101 | Val   ] MPJPE loss: 91.005 | KL loss: 0.750 | Total loss: 91.756\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:53,  2.06it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 13 | Iteration:     1 | Train ] MPJPE loss: 89.496 | KL loss: 0.750 | Total loss: 90.247\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:38,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 13 | Iteration:   101 | Train ] MPJPE loss: 88.839 | KL loss: 0.748 | Total loss: 89.587\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:15,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 13 | Iteration:   201 | Train ] MPJPE loss: 82.965 | KL loss: 0.745 | Total loss: 83.711\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:51,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 13 | Iteration:   301 | Train ] MPJPE loss: 86.438 | KL loss: 0.742 | Total loss: 87.180\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:26,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 13 | Iteration:   401 | Train ] MPJPE loss: 79.708 | KL loss: 0.740 | Total loss: 80.448\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [13:00,  1.57s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 13 | Iteration:   501 | Train ] MPJPE loss: 85.744 | KL loss: 0.737 | Total loss: 86.482\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:34,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 13 | Iteration:   601 | Train ] MPJPE loss: 83.853 | KL loss: 0.735 | Total loss: 84.587\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:08,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 13 | Iteration:   701 | Train ] MPJPE loss: 84.668 | KL loss: 0.732 | Total loss: 85.400\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:14,  1.56s/it]\n",
            "1it [00:00,  1.15it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 13 | Iteration:     1 | Val   ] MPJPE loss: 83.620 | KL loss: 0.732 | Total loss: 84.352\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:49,  2.06it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 13 | Iteration:   101 | Val   ] MPJPE loss: 92.142 | KL loss: 0.732 | Total loss: 92.874\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:54,  2.01it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 14 | Iteration:     1 | Train ] MPJPE loss: 87.681 | KL loss: 0.732 | Total loss: 88.413\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:39,  1.58s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 14 | Iteration:   101 | Train ] MPJPE loss: 90.065 | KL loss: 0.729 | Total loss: 90.794\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:18,  1.57s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 14 | Iteration:   201 | Train ] MPJPE loss: 80.895 | KL loss: 0.727 | Total loss: 81.621\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:54,  1.55s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 14 | Iteration:   301 | Train ] MPJPE loss: 89.538 | KL loss: 0.724 | Total loss: 90.262\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:29,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 14 | Iteration:   401 | Train ] MPJPE loss: 82.498 | KL loss: 0.722 | Total loss: 83.219\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [13:03,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 14 | Iteration:   501 | Train ] MPJPE loss: 85.667 | KL loss: 0.719 | Total loss: 86.386\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:37,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 14 | Iteration:   601 | Train ] MPJPE loss: 81.149 | KL loss: 0.717 | Total loss: 81.865\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:11,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 14 | Iteration:   701 | Train ] MPJPE loss: 82.858 | KL loss: 0.714 | Total loss: 83.572\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:18,  1.56s/it]\n",
            "1it [00:00,  1.16it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 14 | Iteration:     1 | Val   ] MPJPE loss: 84.169 | KL loss: 0.714 | Total loss: 84.882\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:49,  2.09it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 14 | Iteration:   101 | Val   ] MPJPE loss: 92.157 | KL loss: 0.714 | Total loss: 92.870\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:54,  2.01it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 15 | Iteration:     1 | Train ] MPJPE loss: 81.739 | KL loss: 0.714 | Total loss: 82.453\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:39,  1.62s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 15 | Iteration:   101 | Train ] MPJPE loss: 92.280 | KL loss: 0.711 | Total loss: 92.991\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:18,  1.65s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 15 | Iteration:   201 | Train ] MPJPE loss: 84.524 | KL loss: 0.709 | Total loss: 85.233\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:54,  1.55s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 15 | Iteration:   301 | Train ] MPJPE loss: 86.439 | KL loss: 0.706 | Total loss: 87.145\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:30,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 15 | Iteration:   401 | Train ] MPJPE loss: 94.633 | KL loss: 0.704 | Total loss: 95.337\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [13:05,  1.55s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 15 | Iteration:   501 | Train ] MPJPE loss: 87.523 | KL loss: 0.701 | Total loss: 88.224\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:39,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 15 | Iteration:   601 | Train ] MPJPE loss: 82.822 | KL loss: 0.699 | Total loss: 83.520\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:12,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 15 | Iteration:   701 | Train ] MPJPE loss: 85.713 | KL loss: 0.696 | Total loss: 86.409\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:19,  1.56s/it]\n",
            "1it [00:00,  1.13it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 15 | Iteration:     1 | Val   ] MPJPE loss: 85.472 | KL loss: 0.696 | Total loss: 86.167\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:49,  2.04it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 15 | Iteration:   101 | Val   ] MPJPE loss: 94.830 | KL loss: 0.696 | Total loss: 95.526\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:55,  1.99it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 16 | Iteration:     1 | Train ] MPJPE loss: 82.208 | KL loss: 0.696 | Total loss: 82.904\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:38,  1.57s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 16 | Iteration:   101 | Train ] MPJPE loss: 83.319 | KL loss: 0.693 | Total loss: 84.013\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:15,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 16 | Iteration:   201 | Train ] MPJPE loss: 89.523 | KL loss: 0.691 | Total loss: 90.214\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:51,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 16 | Iteration:   301 | Train ] MPJPE loss: 83.333 | KL loss: 0.688 | Total loss: 84.021\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:26,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 16 | Iteration:   401 | Train ] MPJPE loss: 79.902 | KL loss: 0.686 | Total loss: 80.588\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [12:59,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 16 | Iteration:   501 | Train ] MPJPE loss: 84.436 | KL loss: 0.683 | Total loss: 85.119\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:32,  1.52s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 16 | Iteration:   601 | Train ] MPJPE loss: 82.861 | KL loss: 0.681 | Total loss: 83.542\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:05,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 16 | Iteration:   701 | Train ] MPJPE loss: 85.937 | KL loss: 0.678 | Total loss: 86.615\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:12,  1.55s/it]\n",
            "1it [00:00,  1.20it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 16 | Iteration:     1 | Val   ] MPJPE loss: 83.903 | KL loss: 0.678 | Total loss: 84.581\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:47,  2.17it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 16 | Iteration:   101 | Val   ] MPJPE loss: 94.287 | KL loss: 0.678 | Total loss: 94.965\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:51,  2.12it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 17 | Iteration:     1 | Train ] MPJPE loss: 81.418 | KL loss: 0.678 | Total loss: 82.097\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:36,  1.62s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 17 | Iteration:   101 | Train ] MPJPE loss: 80.964 | KL loss: 0.676 | Total loss: 81.640\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:12,  1.57s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 17 | Iteration:   201 | Train ] MPJPE loss: 75.989 | KL loss: 0.673 | Total loss: 76.662\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:46,  1.52s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 17 | Iteration:   301 | Train ] MPJPE loss: 81.815 | KL loss: 0.671 | Total loss: 82.485\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:19,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 17 | Iteration:   401 | Train ] MPJPE loss: 79.489 | KL loss: 0.668 | Total loss: 80.157\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [12:51,  1.51s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 17 | Iteration:   501 | Train ] MPJPE loss: 87.470 | KL loss: 0.666 | Total loss: 88.136\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:24,  1.51s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 17 | Iteration:   601 | Train ] MPJPE loss: 82.695 | KL loss: 0.663 | Total loss: 83.358\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [17:56,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 17 | Iteration:   701 | Train ] MPJPE loss: 79.338 | KL loss: 0.661 | Total loss: 79.999\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:02,  1.54s/it]\n",
            "1it [00:00,  1.16it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 17 | Iteration:     1 | Val   ] MPJPE loss: 80.134 | KL loss: 0.660 | Total loss: 80.795\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:47,  2.15it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 17 | Iteration:   101 | Val   ] MPJPE loss: 92.855 | KL loss: 0.660 | Total loss: 93.516\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:52,  2.08it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 18 | Iteration:     1 | Train ] MPJPE loss: 84.316 | KL loss: 0.660 | Total loss: 84.977\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:36,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 18 | Iteration:   101 | Train ] MPJPE loss: 86.628 | KL loss: 0.658 | Total loss: 87.286\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:12,  1.55s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 18 | Iteration:   201 | Train ] MPJPE loss: 84.788 | KL loss: 0.656 | Total loss: 85.444\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:46,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 18 | Iteration:   301 | Train ] MPJPE loss: 86.028 | KL loss: 0.653 | Total loss: 86.681\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:20,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 18 | Iteration:   401 | Train ] MPJPE loss: 78.736 | KL loss: 0.651 | Total loss: 79.387\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [12:52,  1.52s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 18 | Iteration:   501 | Train ] MPJPE loss: 75.924 | KL loss: 0.648 | Total loss: 76.573\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:25,  1.51s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 18 | Iteration:   601 | Train ] MPJPE loss: 89.019 | KL loss: 0.646 | Total loss: 89.665\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [17:57,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 18 | Iteration:   701 | Train ] MPJPE loss: 86.898 | KL loss: 0.643 | Total loss: 87.541\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:04,  1.54s/it]\n",
            "1it [00:00,  1.15it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 18 | Iteration:     1 | Val   ] MPJPE loss: 80.989 | KL loss: 0.643 | Total loss: 81.632\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:48,  2.14it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 18 | Iteration:   101 | Val   ] MPJPE loss: 94.973 | KL loss: 0.643 | Total loss: 95.616\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:53,  2.06it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 19 | Iteration:     1 | Train ] MPJPE loss: 88.407 | KL loss: 0.643 | Total loss: 89.050\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:38,  1.61s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 19 | Iteration:   101 | Train ] MPJPE loss: 80.980 | KL loss: 0.641 | Total loss: 81.620\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:14,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 19 | Iteration:   201 | Train ] MPJPE loss: 83.290 | KL loss: 0.638 | Total loss: 83.929\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:49,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 19 | Iteration:   301 | Train ] MPJPE loss: 84.411 | KL loss: 0.636 | Total loss: 85.047\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:23,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 19 | Iteration:   401 | Train ] MPJPE loss: 79.062 | KL loss: 0.633 | Total loss: 79.695\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [12:56,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 19 | Iteration:   501 | Train ] MPJPE loss: 83.623 | KL loss: 0.631 | Total loss: 84.254\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:29,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 19 | Iteration:   601 | Train ] MPJPE loss: 80.856 | KL loss: 0.629 | Total loss: 81.485\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:02,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 19 | Iteration:   701 | Train ] MPJPE loss: 82.573 | KL loss: 0.626 | Total loss: 83.199\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:09,  1.55s/it]\n",
            "1it [00:00,  1.16it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 19 | Iteration:     1 | Val   ] MPJPE loss: 79.672 | KL loss: 0.626 | Total loss: 80.298\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:48,  2.13it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 19 | Iteration:   101 | Val   ] MPJPE loss: 99.308 | KL loss: 0.626 | Total loss: 99.935\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:53,  2.05it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 20 | Iteration:     1 | Train ] MPJPE loss: 81.304 | KL loss: 0.626 | Total loss: 81.930\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:37,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 20 | Iteration:   101 | Train ] MPJPE loss: 81.331 | KL loss: 0.624 | Total loss: 81.954\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:14,  1.55s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 20 | Iteration:   201 | Train ] MPJPE loss: 80.463 | KL loss: 0.621 | Total loss: 81.084\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:50,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 20 | Iteration:   301 | Train ] MPJPE loss: 87.509 | KL loss: 0.619 | Total loss: 88.128\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:24,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 20 | Iteration:   401 | Train ] MPJPE loss: 81.381 | KL loss: 0.617 | Total loss: 81.998\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [12:58,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 20 | Iteration:   501 | Train ] MPJPE loss: 84.863 | KL loss: 0.614 | Total loss: 85.478\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:31,  1.51s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 20 | Iteration:   601 | Train ] MPJPE loss: 81.393 | KL loss: 0.612 | Total loss: 82.005\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:04,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 20 | Iteration:   701 | Train ] MPJPE loss: 77.358 | KL loss: 0.610 | Total loss: 77.968\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:11,  1.55s/it]\n",
            "1it [00:00,  1.15it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 20 | Iteration:     1 | Val   ] MPJPE loss: 79.416 | KL loss: 0.609 | Total loss: 80.026\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:48,  2.12it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 20 | Iteration:   101 | Val   ] MPJPE loss: 99.786 | KL loss: 0.609 | Total loss: 100.395\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:52,  2.08it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 21 | Iteration:     1 | Train ] MPJPE loss: 86.685 | KL loss: 0.609 | Total loss: 87.295\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:39,  1.64s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 21 | Iteration:   101 | Train ] MPJPE loss: 76.286 | KL loss: 0.607 | Total loss: 76.893\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:16,  1.59s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 21 | Iteration:   201 | Train ] MPJPE loss: 80.067 | KL loss: 0.605 | Total loss: 80.671\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:52,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 21 | Iteration:   301 | Train ] MPJPE loss: 78.508 | KL loss: 0.602 | Total loss: 79.111\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:26,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 21 | Iteration:   401 | Train ] MPJPE loss: 82.780 | KL loss: 0.600 | Total loss: 83.380\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [13:00,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 21 | Iteration:   501 | Train ] MPJPE loss: 78.214 | KL loss: 0.598 | Total loss: 78.812\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:33,  1.52s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 21 | Iteration:   601 | Train ] MPJPE loss: 81.427 | KL loss: 0.595 | Total loss: 82.023\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:06,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 21 | Iteration:   701 | Train ] MPJPE loss: 89.117 | KL loss: 0.593 | Total loss: 89.710\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:13,  1.55s/it]\n",
            "1it [00:00,  1.17it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 21 | Iteration:     1 | Val   ] MPJPE loss: 77.858 | KL loss: 0.593 | Total loss: 78.451\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:48,  2.09it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 21 | Iteration:   101 | Val   ] MPJPE loss: 97.834 | KL loss: 0.593 | Total loss: 98.427\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:53,  2.05it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 22 | Iteration:     1 | Train ] MPJPE loss: 78.518 | KL loss: 0.593 | Total loss: 79.111\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:38,  1.57s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 22 | Iteration:   101 | Train ] MPJPE loss: 84.901 | KL loss: 0.591 | Total loss: 85.491\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:15,  1.56s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 22 | Iteration:   201 | Train ] MPJPE loss: 82.137 | KL loss: 0.588 | Total loss: 82.726\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:50,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 22 | Iteration:   301 | Train ] MPJPE loss: 76.596 | KL loss: 0.586 | Total loss: 77.182\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:24,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 22 | Iteration:   401 | Train ] MPJPE loss: 78.364 | KL loss: 0.584 | Total loss: 78.948\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [12:58,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 22 | Iteration:   501 | Train ] MPJPE loss: 77.132 | KL loss: 0.582 | Total loss: 77.714\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:31,  1.52s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 22 | Iteration:   601 | Train ] MPJPE loss: 84.491 | KL loss: 0.579 | Total loss: 85.070\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:04,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 22 | Iteration:   701 | Train ] MPJPE loss: 77.030 | KL loss: 0.577 | Total loss: 77.607\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:11,  1.55s/it]\n",
            "1it [00:00,  1.14it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 22 | Iteration:     1 | Val   ] MPJPE loss: 80.845 | KL loss: 0.577 | Total loss: 81.422\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:48,  2.12it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 22 | Iteration:   101 | Val   ] MPJPE loss: 99.363 | KL loss: 0.577 | Total loss: 99.940\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:53,  2.06it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 23 | Iteration:     1 | Train ] MPJPE loss: 82.253 | KL loss: 0.577 | Total loss: 82.830\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:38,  1.57s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 23 | Iteration:   101 | Train ] MPJPE loss: 82.114 | KL loss: 0.575 | Total loss: 82.689\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:15,  1.61s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 23 | Iteration:   201 | Train ] MPJPE loss: 84.113 | KL loss: 0.572 | Total loss: 84.685\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:50,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 23 | Iteration:   301 | Train ] MPJPE loss: 77.347 | KL loss: 0.570 | Total loss: 77.917\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:24,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 23 | Iteration:   401 | Train ] MPJPE loss: 83.124 | KL loss: 0.568 | Total loss: 83.692\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [12:58,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 23 | Iteration:   501 | Train ] MPJPE loss: 80.682 | KL loss: 0.566 | Total loss: 81.248\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:31,  1.52s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 23 | Iteration:   601 | Train ] MPJPE loss: 81.937 | KL loss: 0.564 | Total loss: 82.501\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:05,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 23 | Iteration:   701 | Train ] MPJPE loss: 77.086 | KL loss: 0.561 | Total loss: 77.648\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:12,  1.55s/it]\n",
            "1it [00:00,  1.19it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 23 | Iteration:     1 | Val   ] MPJPE loss: 83.560 | KL loss: 0.561 | Total loss: 84.122\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:48,  2.12it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 23 | Iteration:   101 | Val   ] MPJPE loss: 99.211 | KL loss: 0.561 | Total loss: 99.772\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:53,  2.05it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 24 | Iteration:     1 | Train ] MPJPE loss: 83.828 | KL loss: 0.561 | Total loss: 84.389\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:38,  1.59s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 24 | Iteration:   101 | Train ] MPJPE loss: 77.334 | KL loss: 0.559 | Total loss: 77.893\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:15,  1.57s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 24 | Iteration:   201 | Train ] MPJPE loss: 82.995 | KL loss: 0.557 | Total loss: 83.552\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:50,  1.59s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 24 | Iteration:   301 | Train ] MPJPE loss: 78.114 | KL loss: 0.555 | Total loss: 78.669\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:23,  1.52s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 24 | Iteration:   401 | Train ] MPJPE loss: 80.131 | KL loss: 0.552 | Total loss: 80.683\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [12:55,  1.52s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 24 | Iteration:   501 | Train ] MPJPE loss: 81.555 | KL loss: 0.550 | Total loss: 82.105\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:28,  1.52s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 24 | Iteration:   601 | Train ] MPJPE loss: 83.988 | KL loss: 0.548 | Total loss: 84.536\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [18:00,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 24 | Iteration:   701 | Train ] MPJPE loss: 82.006 | KL loss: 0.546 | Total loss: 82.552\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:07,  1.54s/it]\n",
            "1it [00:00,  1.09it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 24 | Iteration:     1 | Val   ] MPJPE loss: 79.297 | KL loss: 0.546 | Total loss: 79.843\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:47,  2.12it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 24 | Iteration:   101 | Val   ] MPJPE loss: 100.298 | KL loss: 0.546 | Total loss: 100.844\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:52,  2.08it/s]\n",
            "0it [00:00, ?it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 25 | Iteration:     1 | Train ] MPJPE loss: 81.021 | KL loss: 0.546 | Total loss: 81.567\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100it [02:37,  1.57s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 25 | Iteration:   101 | Train ] MPJPE loss: 74.253 | KL loss: 0.544 | Total loss: 74.796\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "200it [05:13,  1.55s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 25 | Iteration:   201 | Train ] MPJPE loss: 74.372 | KL loss: 0.542 | Total loss: 74.913\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "300it [07:48,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 25 | Iteration:   301 | Train ] MPJPE loss: 74.357 | KL loss: 0.539 | Total loss: 74.896\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "400it [10:21,  1.53s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 25 | Iteration:   401 | Train ] MPJPE loss: 83.523 | KL loss: 0.537 | Total loss: 84.060\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "500it [12:53,  1.51s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 25 | Iteration:   501 | Train ] MPJPE loss: 82.603 | KL loss: 0.535 | Total loss: 83.138\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "600it [15:26,  1.51s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 25 | Iteration:   601 | Train ] MPJPE loss: 78.634 | KL loss: 0.533 | Total loss: 79.167\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "700it [17:58,  1.54s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Epoch: 25 | Iteration:   701 | Train ] MPJPE loss: 82.102 | KL loss: 0.531 | Total loss: 82.633\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "704it [18:05,  1.54s/it]\n",
            "1it [00:00,  1.17it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 25 | Iteration:     1 | Val   ] MPJPE loss: 76.538 | KL loss: 0.531 | Total loss: 77.068\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "101it [00:47,  2.16it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m[Epoch: 25 | Iteration:   101 | Val   ] MPJPE loss: 100.416 | KL loss: 0.531 | Total loss: 100.947\u001b[0m\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "110it [00:52,  2.08it/s]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "train_losses, val_losses = model_training.train(data_loader=train_dataset,\n",
        "                                                vald_loader=vald_dataset)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Testing"
      ],
      "metadata": {
        "id": "DPkykXZQiNSS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "_, _, test_dataset = load_dataset()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nUEchQ8EDPrW",
        "outputId": "23bee1f9-b31f-4ec9-bfa9-1745d407a884"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Loading Train Dataset...\n",
            "Loading Validation Dataset...\n",
            "Loading Test Dataset...\n",
            ">>> Training dataset length: 180077\n",
            ">>> Validation dataset length: 28110\n",
            ">>> Test dataset length: 3840\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "num_channels = 3\n",
        "output_n = 25\n",
        "input_n = 10\n",
        "num_joints = 22\n",
        "num_heads = 8\n",
        "dropout = 0.3\n",
        "d_model = 512\n",
        "num_predictions = 100\n",
        "\n",
        "model = Model(num_channels=num_channels,\n",
        "              num_frames_out=output_n,\n",
        "              old_frames=input_n,\n",
        "              num_joints=num_joints,\n",
        "              num_heads=num_heads,\n",
        "              drop=dropout,\n",
        "              d_model=d_model,\n",
        "              num_predictions=num_predictions)"
      ],
      "metadata": {
        "id": "3jGs3taJ4D-r"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "ckpt_path = path\n",
        "output_n = 25\n",
        "input_n = 10\n",
        "actions_to_consider_test = 'all'\n",
        "\n",
        "test(model=model,\n",
        "     ckpt_path=ckpt_path,\n",
        "     test_loader=test_dataset,\n",
        "     input_n=input_n,\n",
        "     output_n=output_n,\n",
        "     actions_to_consider_test=actions_to_consider_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0ijLXf4LD3Gy",
        "outputId": "300bc1be-545a-4087-cc83-4beb5d88fed1"
      },
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model loaded\n",
            "walking :  73.5\n",
            "eating :  63.0\n",
            "smoking :  61.1\n",
            "discussion :  98.7\n",
            "directions :  87.8\n",
            "greeting :  117.7\n",
            "phoning :  79.7\n",
            "posing :  120.6\n",
            "purchases :  107.2\n",
            "sitting :  84.3\n",
            "sittingdown :  107.1\n",
            "takingphoto :  83.9\n",
            "waiting :  88.9\n",
            "walkingdog :  117.1\n",
            "walkingtogether :  65.7\n",
            "Average: 90.4\n",
            "Prediction time:  3.8380204730563694\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "num_channels = 3\n",
        "output_n = 25\n",
        "input_n = 10\n",
        "num_joints = 22\n",
        "num_heads = 8\n",
        "dropout = 0.3\n",
        "d_model = 512\n",
        "num_predictions = 200\n",
        "return_sample = True\n",
        "\n",
        "model = Model(num_channels=num_channels,\n",
        "              num_frames_out=output_n,\n",
        "              old_frames=input_n,\n",
        "              num_joints=num_joints,\n",
        "              num_heads=num_heads,\n",
        "              drop=dropout,\n",
        "              d_model=d_model,\n",
        "              num_predictions=num_predictions,\n",
        "              return_sample=return_sample)"
      ],
      "metadata": {
        "id": "hLf4m50iMUOT"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "mu, ci, isin, dev = get_ci(model=model,\n",
        "                           ckpt_path=ckpt_path,\n",
        "                           test_loader=test_dataset,\n",
        "                           input_n=input_n,\n",
        "                           output_n=output_n)"
      ],
      "metadata": {
        "id": "bfqk4ibnjbS1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plot_ci(ci, cmap='RdPu')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 564
        },
        "outputId": "b0b16721-bd14-4923-96d0-81701c8487d1",
        "id": "vF_lhk8giUjs"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x600 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plot_loss_conf(ci, dev)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 476
        },
        "id": "yVIVL51wMYNX",
        "outputId": "f164d515-608e-4d5d-ab84-cd160055c4f9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    }
  ]
}