model/prototyping_neural_surface_reconstruction.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.nn.functional as F\n",
    "import torch\n",
    "import numpy as np\n",
    "from pathlib import Path\n",
    "import plotly.graph_objs as go\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "\n",
    "from initialization import NeuralSurfaceReconstruction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "703"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nsr = NeuralSurfaceReconstruction(\n",
    "    data=Path(\"../data/calculator\"),\n",
    "    device=torch.device(\"cuda:0\"),\n",
    "    image_scale=1/32,\n",
    ")\n",
    "len(nsr.images)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "optimizer = torch.optim.Adam(nsr.model.parameters(), lr=0.01, eps=10e-12)\n",
    "# scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(optimizer, mode=\"min\",\n",
    "#                                                        patience=10, factor=0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_batched_viewpoint(total: int, step: int, batch_size: int, epochs: int):\n",
    "    return np.tile(\n",
    "        np.arange(\n",
    "            0, total, step\n",
    "        )[:(int((total/step)//batch_size)*batch_size)].reshape((-1, batch_size))[np.newaxis],\n",
    "        (epochs, 1, 1)\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[  0,  10,  20,  30,  40,  50],\n",
       "        [ 60,  70,  80,  90, 100, 110],\n",
       "        [120, 130, 140, 150, 160, 170],\n",
       "        [180, 190, 200, 210, 220, 230],\n",
       "        [240, 250, 260, 270, 280, 290],\n",
       "        [300, 310, 320, 330, 340, 350],\n",
       "        [360, 370, 380, 390, 400, 410],\n",
       "        [420, 430, 440, 450, 460, 470],\n",
       "        [480, 490, 500, 510, 520, 530],\n",
       "        [540, 550, 560, 570, 580, 590],\n",
       "        [600, 610, 620, 630, 640, 650]]])"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "epochs = get_batched_viewpoint(total=len(nsr.images),\n",
    "                               step=10,\n",
    "                               batch_size=6, \n",
    "                               epochs=1)\n",
    "epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "EPOCH: 0 SHARPNESS: 1 FG VOXEL SIZE: 16\n",
      "BATCH: 0\n",
      "::: PREPROCESSING: 0\n",
      "::: PREPROCESSING: 20\n",
      "::: PREPROCESSING: 40\n",
      "::: PREPROCESSING: 60\n",
      "::: PREPROCESSING: 80\n",
      "::: PREPROCESSING: 100\n",
      "::: PREPROCESSING: 120\n",
      "::: PREPROCESSING: 140\n",
      "::: PREPROCESSING: 160\n",
      "::: PREPROCESSING: 180\n",
      "::: PREPROCESSING: 200\n",
      "::: PREPROCESSING: 220\n",
      "::: PREPROCESSING: 240\n",
      "::: PREPROCESSING: 260\n",
      "::: PREPROCESSING: 280\n",
      "::: PREPROCESSING: 300\n",
      "::: PREPROCESSING: 320\n",
      "::: PREPROCESSING: 340\n",
      "::: PREPROCESSING: 360\n",
      "::: PREPROCESSING: 380\n",
      "::: PREPROCESSING: 400\n",
      "::: PREPROCESSING: 420\n",
      "::: PREPROCESSING: 440\n",
      "::: PREPROCESSING: 460\n",
      "IMAGE EPOCH: 0 LOSS:  151501.828125\n",
      "IMAGE EPOCH: 1 LOSS:  106422.9765625\n",
      "IMAGE EPOCH: 2 LOSS:  80494.9296875\n",
      "IMAGE EPOCH: 3 LOSS:  70070.7578125\n",
      "IMAGE EPOCH: 4 LOSS:  69732.9765625\n",
      "IMAGE EPOCH: 5 LOSS:  70700.59375\n",
      "IMAGE EPOCH: 6 LOSS:  68067.7734375\n",
      "IMAGE EPOCH: 7 LOSS:  62183.8203125\n",
      "IMAGE EPOCH: 8 LOSS:  55354.03515625\n",
      "IMAGE EPOCH: 9 LOSS:  49557.3515625\n",
      "IMAGE EPOCH: 10 LOSS:  45523.8359375\n",
      "IMAGE EPOCH: 11 LOSS:  42922.6015625\n",
      "IMAGE EPOCH: 12 LOSS:  40997.87109375\n",
      "IMAGE EPOCH: 13 LOSS:  39081.890625\n",
      "IMAGE EPOCH: 14 LOSS:  36816.3359375\n",
      "IMAGE EPOCH: 15 LOSS:  34174.453125\n",
      "IMAGE EPOCH: 16 LOSS:  31389.158203125\n",
      "IMAGE EPOCH: 17 LOSS:  28860.10546875\n",
      "IMAGE EPOCH: 18 LOSS:  26983.09375\n",
      "IMAGE EPOCH: 19 LOSS:  25893.708984375\n",
      "IMAGE EPOCH: 20 LOSS:  25307.240234375\n",
      "IMAGE EPOCH: 21 LOSS:  24676.650390625\n",
      "IMAGE EPOCH: 22 LOSS:  23587.703125\n",
      "IMAGE EPOCH: 23 LOSS:  22070.16015625\n",
      "IMAGE EPOCH: 24 LOSS:  20502.3671875\n",
      "IMAGE EPOCH: 25 LOSS:  19252.23046875\n",
      "IMAGE EPOCH: 26 LOSS:  18411.248046875\n",
      "IMAGE EPOCH: 27 LOSS:  17787.01171875\n",
      "IMAGE EPOCH: 28 LOSS:  17135.15625\n",
      "IMAGE EPOCH: 29 LOSS:  16350.03515625\n",
      "IMAGE EPOCH: 30 LOSS:  15494.8251953125\n",
      "IMAGE EPOCH: 31 LOSS:  14763.93359375\n",
      "IMAGE EPOCH: 32 LOSS:  14289.1845703125\n",
      "IMAGE EPOCH: 33 LOSS:  13974.7177734375\n",
      "IMAGE EPOCH: 34 LOSS:  13563.55078125\n",
      "IMAGE EPOCH: 35 LOSS:  12961.1865234375\n",
      "IMAGE EPOCH: 36 LOSS:  12365.892578125\n",
      "IMAGE EPOCH: 37 LOSS:  12004.4130859375\n",
      "IMAGE EPOCH: 38 LOSS:  11856.4033203125\n",
      "IMAGE EPOCH: 39 LOSS:  11697.890625\n",
      "IMAGE EPOCH: 40 LOSS:  11368.2958984375\n",
      "IMAGE EPOCH: 41 LOSS:  10931.0693359375\n",
      "IMAGE EPOCH: 42 LOSS:  10583.2294921875\n",
      "IMAGE EPOCH: 43 LOSS:  10426.9775390625\n",
      "IMAGE EPOCH: 44 LOSS:  10349.384765625\n",
      "IMAGE EPOCH: 45 LOSS:  10177.0732421875\n",
      "IMAGE EPOCH: 46 LOSS:  9903.517578125\n",
      "IMAGE EPOCH: 47 LOSS:  9658.822265625\n",
      "IMAGE EPOCH: 48 LOSS:  9511.060546875\n",
      "IMAGE EPOCH: 49 LOSS:  9402.4404296875\n",
      "IMAGE EPOCH: 50 LOSS:  9263.787109375\n",
      "IMAGE EPOCH: 51 LOSS:  9106.181640625\n",
      "IMAGE EPOCH: 52 LOSS:  8972.029296875\n",
      "IMAGE EPOCH: 53 LOSS:  8853.67578125\n",
      "IMAGE EPOCH: 54 LOSS:  8719.0537109375\n",
      "IMAGE EPOCH: 55 LOSS:  8582.3251953125\n",
      "IMAGE EPOCH: 56 LOSS:  8483.8818359375\n",
      "IMAGE EPOCH: 57 LOSS:  8416.537109375\n",
      "IMAGE EPOCH: 58 LOSS:  8331.408203125\n",
      "IMAGE EPOCH: 59 LOSS:  8214.353515625\n",
      "IMAGE EPOCH: 60 LOSS:  8106.30712890625\n",
      "IMAGE EPOCH: 61 LOSS:  8038.8916015625\n",
      "IMAGE EPOCH: 62 LOSS:  7986.6611328125\n",
      "IMAGE EPOCH: 63 LOSS:  7911.25048828125\n",
      "IMAGE EPOCH: 64 LOSS:  7822.01171875\n",
      "IMAGE EPOCH: 65 LOSS:  7751.98876953125\n",
      "IMAGE EPOCH: 66 LOSS:  7701.9033203125\n",
      "IMAGE EPOCH: 67 LOSS:  7647.1689453125\n",
      "IMAGE EPOCH: 68 LOSS:  7581.84912109375\n",
      "IMAGE EPOCH: 69 LOSS:  7521.33154296875\n",
      "IMAGE EPOCH: 70 LOSS:  7470.63232421875\n",
      "IMAGE EPOCH: 71 LOSS:  7419.70947265625\n",
      "IMAGE EPOCH: 72 LOSS:  7366.37646484375\n",
      "IMAGE EPOCH: 73 LOSS:  7318.06689453125\n",
      "IMAGE EPOCH: 74 LOSS:  7274.0244140625\n",
      "IMAGE EPOCH: 75 LOSS:  7227.58837890625\n",
      "IMAGE EPOCH: 76 LOSS:  7181.470703125\n",
      "IMAGE EPOCH: 77 LOSS:  7142.32275390625\n",
      "IMAGE EPOCH: 78 LOSS:  7105.82666015625\n",
      "IMAGE EPOCH: 79 LOSS:  7064.29150390625\n",
      "IMAGE EPOCH: 80 LOSS:  7022.00634765625\n",
      "IMAGE EPOCH: 81 LOSS:  6986.87158203125\n",
      "IMAGE EPOCH: 82 LOSS:  6955.4912109375\n",
      "IMAGE EPOCH: 83 LOSS:  6920.66552734375\n",
      "IMAGE EPOCH: 84 LOSS:  6884.3310546875\n",
      "IMAGE EPOCH: 85 LOSS:  6851.77783203125\n",
      "IMAGE EPOCH: 86 LOSS:  6821.6943359375\n",
      "IMAGE EPOCH: 87 LOSS:  6790.88671875\n",
      "IMAGE EPOCH: 88 LOSS:  6760.40185546875\n",
      "IMAGE EPOCH: 89 LOSS:  6731.28515625\n",
      "IMAGE EPOCH: 90 LOSS:  6702.3662109375\n",
      "IMAGE EPOCH: 91 LOSS:  6674.234375\n",
      "IMAGE EPOCH: 92 LOSS:  6648.1630859375\n",
      "IMAGE EPOCH: 93 LOSS:  6622.349609375\n",
      "IMAGE EPOCH: 94 LOSS:  6595.26025390625\n",
      "IMAGE EPOCH: 95 LOSS:  6569.01611328125\n",
      "IMAGE EPOCH: 96 LOSS:  6544.9130859375\n",
      "IMAGE EPOCH: 97 LOSS:  6520.8876953125\n",
      "IMAGE EPOCH: 98 LOSS:  6495.6796875\n",
      "IMAGE EPOCH: 99 LOSS:  6470.6337890625\n",
      "EPOCH: 1 SHARPNESS: 2.0 FG VOXEL SIZE: 32\n",
      "BATCH: 0\n",
      "::: PREPROCESSING: 0\n",
      "::: PREPROCESSING: 10\n",
      "::: PREPROCESSING: 20\n",
      "::: PREPROCESSING: 30\n",
      "::: PREPROCESSING: 40\n",
      "::: PREPROCESSING: 50\n",
      "::: PREPROCESSING: 60\n",
      "::: PREPROCESSING: 70\n",
      "::: PREPROCESSING: 80\n",
      "::: PREPROCESSING: 90\n",
      "::: PREPROCESSING: 100\n",
      "::: PREPROCESSING: 110\n",
      "IMAGE EPOCH: 0 LOSS:  6016.888671875\n",
      "IMAGE EPOCH: 1 LOSS:  5720.53515625\n",
      "IMAGE EPOCH: 2 LOSS:  5454.4345703125\n",
      "IMAGE EPOCH: 3 LOSS:  5450.18408203125\n",
      "IMAGE EPOCH: 4 LOSS:  5621.2509765625\n",
      "IMAGE EPOCH: 5 LOSS:  5685.67822265625\n",
      "IMAGE EPOCH: 6 LOSS:  5535.66943359375\n",
      "IMAGE EPOCH: 7 LOSS:  5344.25048828125\n",
      "IMAGE EPOCH: 8 LOSS:  5312.7216796875\n",
      "IMAGE EPOCH: 9 LOSS:  5423.2548828125\n",
      "IMAGE EPOCH: 10 LOSS:  5472.0888671875\n",
      "IMAGE EPOCH: 11 LOSS:  5379.4296875\n",
      "IMAGE EPOCH: 12 LOSS:  5263.70068359375\n",
      "IMAGE EPOCH: 13 LOSS:  5246.2265625\n",
      "IMAGE EPOCH: 14 LOSS:  5305.78369140625\n",
      "IMAGE EPOCH: 15 LOSS:  5329.08349609375\n",
      "IMAGE EPOCH: 16 LOSS:  5269.5146484375\n",
      "IMAGE EPOCH: 17 LOSS:  5194.54638671875\n",
      "IMAGE EPOCH: 18 LOSS:  5179.40771484375\n",
      "IMAGE EPOCH: 19 LOSS:  5211.8876953125\n",
      "IMAGE EPOCH: 20 LOSS:  5224.19091796875\n",
      "IMAGE EPOCH: 21 LOSS:  5187.14892578125\n",
      "IMAGE EPOCH: 22 LOSS:  5138.5419921875\n",
      "IMAGE EPOCH: 23 LOSS:  5124.13671875\n",
      "IMAGE EPOCH: 24 LOSS:  5139.0205078125\n",
      "IMAGE EPOCH: 25 LOSS:  5142.208984375\n",
      "IMAGE EPOCH: 26 LOSS:  5116.18505859375\n",
      "IMAGE EPOCH: 27 LOSS:  5086.04541015625\n",
      "IMAGE EPOCH: 28 LOSS:  5078.8740234375\n",
      "IMAGE EPOCH: 29 LOSS:  5087.6962890625\n",
      "IMAGE EPOCH: 30 LOSS:  5086.5439453125\n",
      "IMAGE EPOCH: 31 LOSS:  5067.6240234375\n",
      "IMAGE EPOCH: 32 LOSS:  5048.35791015625\n",
      "IMAGE EPOCH: 33 LOSS:  5044.19580078125\n",
      "IMAGE EPOCH: 34 LOSS:  5048.23095703125\n",
      "IMAGE EPOCH: 35 LOSS:  5044.12646484375\n",
      "IMAGE EPOCH: 36 LOSS:  5030.10302734375\n",
      "IMAGE EPOCH: 37 LOSS:  5018.9111328125\n",
      "IMAGE EPOCH: 38 LOSS:  5017.57861328125\n",
      "IMAGE EPOCH: 39 LOSS:  5018.58203125\n",
      "IMAGE EPOCH: 40 LOSS:  5012.75439453125\n",
      "IMAGE EPOCH: 41 LOSS:  5002.3125\n",
      "IMAGE EPOCH: 42 LOSS:  4995.84228515625\n",
      "IMAGE EPOCH: 43 LOSS:  4995.14990234375\n",
      "IMAGE EPOCH: 44 LOSS:  4993.78662109375\n",
      "IMAGE EPOCH: 45 LOSS:  4987.6435546875\n",
      "IMAGE EPOCH: 46 LOSS:  4980.43505859375\n",
      "IMAGE EPOCH: 47 LOSS:  4976.89697265625\n",
      "IMAGE EPOCH: 48 LOSS:  4975.73095703125\n",
      "IMAGE EPOCH: 49 LOSS:  4972.58056640625\n",
      "IMAGE EPOCH: 50 LOSS:  4966.91845703125\n",
      "IMAGE EPOCH: 51 LOSS:  4962.13525390625\n",
      "IMAGE EPOCH: 52 LOSS:  4959.849609375\n",
      "IMAGE EPOCH: 53 LOSS:  4957.787109375\n",
      "IMAGE EPOCH: 54 LOSS:  4953.83642578125\n",
      "IMAGE EPOCH: 55 LOSS:  4949.1943359375\n",
      "IMAGE EPOCH: 56 LOSS:  4945.9951171875\n",
      "IMAGE EPOCH: 57 LOSS:  4943.94384765625\n",
      "IMAGE EPOCH: 58 LOSS:  4941.18994140625\n",
      "IMAGE EPOCH: 59 LOSS:  4937.38037109375\n",
      "IMAGE EPOCH: 60 LOSS:  4933.91796875\n",
      "IMAGE EPOCH: 61 LOSS:  4931.52001953125\n",
      "IMAGE EPOCH: 62 LOSS:  4929.2392578125\n",
      "IMAGE EPOCH: 63 LOSS:  4926.21923828125\n",
      "IMAGE EPOCH: 64 LOSS:  4922.9990234375\n",
      "IMAGE EPOCH: 65 LOSS:  4920.42578125\n",
      "IMAGE EPOCH: 66 LOSS:  4918.2919921875\n",
      "IMAGE EPOCH: 67 LOSS:  4915.83740234375\n",
      "IMAGE EPOCH: 68 LOSS:  4913.03564453125\n",
      "IMAGE EPOCH: 69 LOSS:  4910.501953125\n",
      "IMAGE EPOCH: 70 LOSS:  4908.4140625\n",
      "IMAGE EPOCH: 71 LOSS:  4906.3056640625\n",
      "IMAGE EPOCH: 72 LOSS:  4903.91943359375\n",
      "IMAGE EPOCH: 73 LOSS:  4901.57373046875\n",
      "IMAGE EPOCH: 74 LOSS:  4899.5498046875\n",
      "IMAGE EPOCH: 75 LOSS:  4897.6533203125\n",
      "IMAGE EPOCH: 76 LOSS:  4895.6083984375\n",
      "IMAGE EPOCH: 77 LOSS:  4893.51171875\n",
      "IMAGE EPOCH: 78 LOSS:  4891.61474609375\n",
      "IMAGE EPOCH: 79 LOSS:  4889.8798828125\n",
      "IMAGE EPOCH: 80 LOSS:  4888.08740234375\n",
      "IMAGE EPOCH: 81 LOSS:  4886.22119140625\n",
      "IMAGE EPOCH: 82 LOSS:  4884.45703125\n",
      "IMAGE EPOCH: 83 LOSS:  4882.8388671875\n",
      "IMAGE EPOCH: 84 LOSS:  4881.23095703125\n",
      "IMAGE EPOCH: 85 LOSS:  4879.57666015625\n",
      "IMAGE EPOCH: 86 LOSS:  4877.97705078125\n",
      "IMAGE EPOCH: 87 LOSS:  4876.4912109375\n",
      "IMAGE EPOCH: 88 LOSS:  4875.044921875\n",
      "IMAGE EPOCH: 89 LOSS:  4873.5771484375\n",
      "IMAGE EPOCH: 90 LOSS:  4872.14208984375\n",
      "IMAGE EPOCH: 91 LOSS:  4870.7958984375\n",
      "IMAGE EPOCH: 92 LOSS:  4869.49853515625\n",
      "IMAGE EPOCH: 93 LOSS:  4868.19677734375\n",
      "IMAGE EPOCH: 94 LOSS:  4866.91455078125\n",
      "IMAGE EPOCH: 95 LOSS:  4865.6953125\n",
      "IMAGE EPOCH: 96 LOSS:  4864.5205078125\n",
      "IMAGE EPOCH: 97 LOSS:  4863.3486328125\n",
      "IMAGE EPOCH: 98 LOSS:  4862.19091796875\n",
      "IMAGE EPOCH: 99 LOSS:  4861.0771484375\n",
      "BATCH: 1\n",
      "::: PREPROCESSING: 120\n",
      "::: PREPROCESSING: 130\n",
      "::: PREPROCESSING: 140\n",
      "::: PREPROCESSING: 150\n",
      "::: PREPROCESSING: 160\n",
      "::: PREPROCESSING: 170\n",
      "::: PREPROCESSING: 180\n",
      "::: PREPROCESSING: 190\n",
      "::: PREPROCESSING: 200\n",
      "::: PREPROCESSING: 210\n",
      "::: PREPROCESSING: 220\n",
      "::: PREPROCESSING: 230\n",
      "IMAGE EPOCH: 0 LOSS:  5914.78076171875\n",
      "IMAGE EPOCH: 1 LOSS:  4818.9873046875\n",
      "IMAGE EPOCH: 2 LOSS:  5192.478515625\n",
      "IMAGE EPOCH: 3 LOSS:  5549.54736328125\n",
      "IMAGE EPOCH: 4 LOSS:  4848.53662109375\n",
      "IMAGE EPOCH: 5 LOSS:  4652.552734375\n",
      "IMAGE EPOCH: 6 LOSS:  5086.357421875\n",
      "IMAGE EPOCH: 7 LOSS:  4984.5966796875\n",
      "IMAGE EPOCH: 8 LOSS:  4554.0185546875\n",
      "IMAGE EPOCH: 9 LOSS:  4602.421875\n",
      "IMAGE EPOCH: 10 LOSS:  4846.1142578125\n",
      "IMAGE EPOCH: 11 LOSS:  4706.28857421875\n",
      "IMAGE EPOCH: 12 LOSS:  4479.3193359375\n",
      "IMAGE EPOCH: 13 LOSS:  4572.10546875\n",
      "IMAGE EPOCH: 14 LOSS:  4723.97900390625\n",
      "IMAGE EPOCH: 15 LOSS:  4627.078125\n",
      "IMAGE EPOCH: 16 LOSS:  4488.80908203125\n",
      "IMAGE EPOCH: 17 LOSS:  4547.00537109375\n",
      "IMAGE EPOCH: 18 LOSS:  4640.060546875\n",
      "IMAGE EPOCH: 19 LOSS:  4576.58251953125\n",
      "IMAGE EPOCH: 20 LOSS:  4482.7236328125\n",
      "IMAGE EPOCH: 21 LOSS:  4511.4052734375\n",
      "IMAGE EPOCH: 22 LOSS:  4567.36376953125\n",
      "IMAGE EPOCH: 23 LOSS:  4525.4521484375\n",
      "IMAGE EPOCH: 24 LOSS:  4457.9072265625\n",
      "IMAGE EPOCH: 25 LOSS:  4470.47509765625\n",
      "IMAGE EPOCH: 26 LOSS:  4508.5537109375\n",
      "IMAGE EPOCH: 27 LOSS:  4482.8740234375\n",
      "IMAGE EPOCH: 28 LOSS:  4436.11572265625\n",
      "IMAGE EPOCH: 29 LOSS:  4442.20068359375\n",
      "IMAGE EPOCH: 30 LOSS:  4468.265625\n",
      "IMAGE EPOCH: 31 LOSS:  4453.93701171875\n",
      "IMAGE EPOCH: 32 LOSS:  4425.19677734375\n",
      "IMAGE EPOCH: 33 LOSS:  4431.548828125\n",
      "IMAGE EPOCH: 34 LOSS:  4451.19580078125\n",
      "IMAGE EPOCH: 35 LOSS:  4442.46142578125\n",
      "IMAGE EPOCH: 36 LOSS:  4423.1875\n",
      "IMAGE EPOCH: 37 LOSS:  4427.14990234375\n",
      "IMAGE EPOCH: 38 LOSS:  4439.5546875\n",
      "IMAGE EPOCH: 39 LOSS:  4433.373046875\n",
      "IMAGE EPOCH: 40 LOSS:  4420.99169921875\n",
      "IMAGE EPOCH: 41 LOSS:  4423.88330078125\n",
      "IMAGE EPOCH: 42 LOSS:  4431.158203125\n",
      "IMAGE EPOCH: 43 LOSS:  4425.42822265625\n",
      "IMAGE EPOCH: 44 LOSS:  4416.21484375\n",
      "IMAGE EPOCH: 45 LOSS:  4417.5615234375\n",
      "IMAGE EPOCH: 46 LOSS:  4421.572265625\n",
      "IMAGE EPOCH: 47 LOSS:  4417.310546875\n",
      "IMAGE EPOCH: 48 LOSS:  4411.5224609375\n",
      "IMAGE EPOCH: 49 LOSS:  4412.7705078125\n",
      "IMAGE EPOCH: 50 LOSS:  4415.1708984375\n",
      "IMAGE EPOCH: 51 LOSS:  4411.9189453125\n",
      "IMAGE EPOCH: 52 LOSS:  4408.1162109375\n",
      "IMAGE EPOCH: 53 LOSS:  4409.150390625\n",
      "IMAGE EPOCH: 54 LOSS:  4410.6875\n",
      "IMAGE EPOCH: 55 LOSS:  4408.48681640625\n",
      "IMAGE EPOCH: 56 LOSS:  4406.162109375\n",
      "IMAGE EPOCH: 57 LOSS:  4406.94140625\n",
      "IMAGE EPOCH: 58 LOSS:  4407.67724609375\n",
      "IMAGE EPOCH: 59 LOSS:  4405.8759765625\n",
      "IMAGE EPOCH: 60 LOSS:  4404.24951171875\n",
      "IMAGE EPOCH: 61 LOSS:  4404.68701171875\n",
      "IMAGE EPOCH: 62 LOSS:  4404.873046875\n",
      "IMAGE EPOCH: 63 LOSS:  4403.40478515625\n",
      "IMAGE EPOCH: 64 LOSS:  4402.25732421875\n",
      "IMAGE EPOCH: 65 LOSS:  4402.453125\n",
      "IMAGE EPOCH: 66 LOSS:  4402.314453125\n",
      "IMAGE EPOCH: 67 LOSS:  4401.150390625\n",
      "IMAGE EPOCH: 68 LOSS:  4400.39306640625\n",
      "IMAGE EPOCH: 69 LOSS:  4400.48095703125\n",
      "IMAGE EPOCH: 70 LOSS:  4400.20166015625\n",
      "IMAGE EPOCH: 71 LOSS:  4399.310546875\n",
      "IMAGE EPOCH: 72 LOSS:  4398.8271484375\n",
      "IMAGE EPOCH: 73 LOSS:  4398.845703125\n",
      "IMAGE EPOCH: 74 LOSS:  4398.51708984375\n",
      "IMAGE EPOCH: 75 LOSS:  4397.84521484375\n",
      "IMAGE EPOCH: 76 LOSS:  4397.5078125\n",
      "IMAGE EPOCH: 77 LOSS:  4397.41650390625\n",
      "IMAGE EPOCH: 78 LOSS:  4397.02392578125\n",
      "IMAGE EPOCH: 79 LOSS:  4396.47412109375\n",
      "IMAGE EPOCH: 80 LOSS:  4396.19921875\n",
      "IMAGE EPOCH: 81 LOSS:  4396.0244140625\n",
      "IMAGE EPOCH: 82 LOSS:  4395.615234375\n",
      "IMAGE EPOCH: 83 LOSS:  4395.158203125\n",
      "IMAGE EPOCH: 84 LOSS:  4394.896484375\n",
      "IMAGE EPOCH: 85 LOSS:  4394.6484375\n",
      "IMAGE EPOCH: 86 LOSS:  4394.2451171875\n",
      "IMAGE EPOCH: 87 LOSS:  4393.8671875\n",
      "IMAGE EPOCH: 88 LOSS:  4393.623046875\n",
      "IMAGE EPOCH: 89 LOSS:  4393.34765625\n",
      "IMAGE EPOCH: 90 LOSS:  4392.97607421875\n",
      "IMAGE EPOCH: 91 LOSS:  4392.65087890625\n",
      "IMAGE EPOCH: 92 LOSS:  4392.4013671875\n",
      "IMAGE EPOCH: 93 LOSS:  4392.107421875\n",
      "IMAGE EPOCH: 94 LOSS:  4391.76416015625\n",
      "IMAGE EPOCH: 95 LOSS:  4391.47412109375\n",
      "IMAGE EPOCH: 96 LOSS:  4391.216796875\n",
      "IMAGE EPOCH: 97 LOSS:  4390.9130859375\n",
      "IMAGE EPOCH: 98 LOSS:  4390.5927734375\n",
      "IMAGE EPOCH: 99 LOSS:  4390.31396484375\n",
      "BATCH: 2\n",
      "::: PREPROCESSING: 240\n",
      "::: PREPROCESSING: 250\n",
      "::: PREPROCESSING: 260\n",
      "::: PREPROCESSING: 270\n",
      "::: PREPROCESSING: 280\n",
      "::: PREPROCESSING: 290\n",
      "::: PREPROCESSING: 300\n",
      "::: PREPROCESSING: 310\n",
      "::: PREPROCESSING: 320\n",
      "::: PREPROCESSING: 330\n",
      "::: PREPROCESSING: 340\n",
      "::: PREPROCESSING: 350\n",
      "IMAGE EPOCH: 0 LOSS:  4878.3876953125\n",
      "IMAGE EPOCH: 1 LOSS:  4856.6669921875\n",
      "IMAGE EPOCH: 2 LOSS:  4829.8388671875\n",
      "IMAGE EPOCH: 3 LOSS:  4812.900390625\n",
      "IMAGE EPOCH: 4 LOSS:  4807.72021484375\n",
      "IMAGE EPOCH: 5 LOSS:  4805.67236328125\n",
      "IMAGE EPOCH: 6 LOSS:  4798.095703125\n",
      "IMAGE EPOCH: 7 LOSS:  4783.3955078125\n",
      "IMAGE EPOCH: 8 LOSS:  4766.099609375\n",
      "IMAGE EPOCH: 9 LOSS:  4751.81103515625\n",
      "IMAGE EPOCH: 10 LOSS:  4743.28662109375\n",
      "IMAGE EPOCH: 11 LOSS:  4739.615234375\n",
      "IMAGE EPOCH: 12 LOSS:  4737.73583984375\n",
      "IMAGE EPOCH: 13 LOSS:  4734.88525390625\n",
      "IMAGE EPOCH: 14 LOSS:  4730.333984375\n",
      "IMAGE EPOCH: 15 LOSS:  4725.3330078125\n",
      "IMAGE EPOCH: 16 LOSS:  4721.57373046875\n",
      "IMAGE EPOCH: 17 LOSS:  4719.767578125\n",
      "IMAGE EPOCH: 18 LOSS:  4719.3525390625\n",
      "IMAGE EPOCH: 19 LOSS:  4719.14892578125\n",
      "IMAGE EPOCH: 20 LOSS:  4718.1123046875\n",
      "IMAGE EPOCH: 21 LOSS:  4715.8115234375\n",
      "IMAGE EPOCH: 22 LOSS:  4712.62255859375\n",
      "IMAGE EPOCH: 23 LOSS:  4709.50341796875\n",
      "IMAGE EPOCH: 24 LOSS:  4707.3095703125\n",
      "IMAGE EPOCH: 25 LOSS:  4706.056640625\n",
      "IMAGE EPOCH: 26 LOSS:  4704.888671875\n",
      "IMAGE EPOCH: 27 LOSS:  4702.86279296875\n",
      "IMAGE EPOCH: 28 LOSS:  4699.833984375\n",
      "IMAGE EPOCH: 29 LOSS:  4696.5537109375\n",
      "IMAGE EPOCH: 30 LOSS:  4693.93505859375\n",
      "IMAGE EPOCH: 31 LOSS:  4692.2607421875\n",
      "IMAGE EPOCH: 32 LOSS:  4691.08349609375\n",
      "IMAGE EPOCH: 33 LOSS:  4689.76806640625\n",
      "IMAGE EPOCH: 34 LOSS:  4688.05517578125\n",
      "IMAGE EPOCH: 35 LOSS:  4686.14599609375\n",
      "IMAGE EPOCH: 36 LOSS:  4684.41455078125\n",
      "IMAGE EPOCH: 37 LOSS:  4683.09228515625\n",
      "IMAGE EPOCH: 38 LOSS:  4682.1376953125\n",
      "IMAGE EPOCH: 39 LOSS:  4681.322265625\n",
      "IMAGE EPOCH: 40 LOSS:  4680.42333984375\n",
      "IMAGE EPOCH: 41 LOSS:  4679.36328125\n",
      "IMAGE EPOCH: 42 LOSS:  4678.228515625\n",
      "IMAGE EPOCH: 43 LOSS:  4677.16162109375\n",
      "IMAGE EPOCH: 44 LOSS:  4676.24462890625\n",
      "IMAGE EPOCH: 45 LOSS:  4675.44189453125\n",
      "IMAGE EPOCH: 46 LOSS:  4674.662109375\n",
      "IMAGE EPOCH: 47 LOSS:  4673.8349609375\n",
      "IMAGE EPOCH: 48 LOSS:  4672.9404296875\n",
      "IMAGE EPOCH: 49 LOSS:  4672.005859375\n",
      "IMAGE EPOCH: 50 LOSS:  4671.09033203125\n",
      "IMAGE EPOCH: 51 LOSS:  4670.2412109375\n",
      "IMAGE EPOCH: 52 LOSS:  4669.458984375\n",
      "IMAGE EPOCH: 53 LOSS:  4668.697265625\n",
      "IMAGE EPOCH: 54 LOSS:  4667.91064453125\n",
      "IMAGE EPOCH: 55 LOSS:  4667.1015625\n",
      "IMAGE EPOCH: 56 LOSS:  4666.3212890625\n",
      "IMAGE EPOCH: 57 LOSS:  4665.6171875\n",
      "IMAGE EPOCH: 58 LOSS:  4664.98681640625\n",
      "IMAGE EPOCH: 59 LOSS:  4664.384765625\n",
      "IMAGE EPOCH: 60 LOSS:  4663.7744140625\n",
      "IMAGE EPOCH: 61 LOSS:  4663.1513671875\n",
      "IMAGE EPOCH: 62 LOSS:  4662.53466796875\n",
      "IMAGE EPOCH: 63 LOSS:  4661.94580078125\n",
      "IMAGE EPOCH: 64 LOSS:  4661.38720703125\n",
      "IMAGE EPOCH: 65 LOSS:  4660.841796875\n",
      "IMAGE EPOCH: 66 LOSS:  4660.29052734375\n",
      "IMAGE EPOCH: 67 LOSS:  4659.72265625\n",
      "IMAGE EPOCH: 68 LOSS:  4659.14501953125\n",
      "IMAGE EPOCH: 69 LOSS:  4658.5693359375\n",
      "IMAGE EPOCH: 70 LOSS:  4658.00537109375\n",
      "IMAGE EPOCH: 71 LOSS:  4657.44970703125\n",
      "IMAGE EPOCH: 72 LOSS:  4656.8974609375\n",
      "IMAGE EPOCH: 73 LOSS:  4656.34228515625\n",
      "IMAGE EPOCH: 74 LOSS:  4655.78466796875\n",
      "IMAGE EPOCH: 75 LOSS:  4655.228515625\n",
      "IMAGE EPOCH: 76 LOSS:  4654.67724609375\n",
      "IMAGE EPOCH: 77 LOSS:  4654.1328125\n",
      "IMAGE EPOCH: 78 LOSS:  4653.59423828125\n",
      "IMAGE EPOCH: 79 LOSS:  4653.0595703125\n",
      "IMAGE EPOCH: 80 LOSS:  4652.5234375\n",
      "IMAGE EPOCH: 81 LOSS:  4651.98828125\n",
      "IMAGE EPOCH: 82 LOSS:  4651.4541015625\n",
      "IMAGE EPOCH: 83 LOSS:  4650.92578125\n",
      "IMAGE EPOCH: 84 LOSS:  4650.4013671875\n",
      "IMAGE EPOCH: 85 LOSS:  4649.8779296875\n",
      "IMAGE EPOCH: 86 LOSS:  4649.35595703125\n",
      "IMAGE EPOCH: 87 LOSS:  4648.8330078125\n",
      "IMAGE EPOCH: 88 LOSS:  4648.31005859375\n",
      "IMAGE EPOCH: 89 LOSS:  4647.7900390625\n",
      "IMAGE EPOCH: 90 LOSS:  4647.27197265625\n",
      "IMAGE EPOCH: 91 LOSS:  4646.7548828125\n",
      "IMAGE EPOCH: 92 LOSS:  4646.2373046875\n",
      "IMAGE EPOCH: 93 LOSS:  4645.71875\n",
      "IMAGE EPOCH: 94 LOSS:  4645.20263671875\n",
      "IMAGE EPOCH: 95 LOSS:  4644.6875\n",
      "IMAGE EPOCH: 96 LOSS:  4644.1748046875\n",
      "IMAGE EPOCH: 97 LOSS:  4643.6630859375\n",
      "IMAGE EPOCH: 98 LOSS:  4643.1513671875\n",
      "IMAGE EPOCH: 99 LOSS:  4642.64111328125\n",
      "BATCH: 3\n",
      "::: PREPROCESSING: 360\n",
      "::: PREPROCESSING: 370\n",
      "::: PREPROCESSING: 380\n",
      "::: PREPROCESSING: 390\n",
      "::: PREPROCESSING: 400\n",
      "::: PREPROCESSING: 410\n",
      "::: PREPROCESSING: 420\n",
      "::: PREPROCESSING: 430\n",
      "::: PREPROCESSING: 440\n",
      "::: PREPROCESSING: 450\n",
      "::: PREPROCESSING: 460\n",
      "::: PREPROCESSING: 470\n",
      "IMAGE EPOCH: 0 LOSS:  5171.77294921875\n",
      "IMAGE EPOCH: 1 LOSS:  3383.566650390625\n",
      "IMAGE EPOCH: 2 LOSS:  4497.8974609375\n",
      "IMAGE EPOCH: 3 LOSS:  4783.408203125\n",
      "IMAGE EPOCH: 4 LOSS:  3492.63232421875\n",
      "IMAGE EPOCH: 5 LOSS:  3790.20751953125\n",
      "IMAGE EPOCH: 6 LOSS:  4407.11083984375\n",
      "IMAGE EPOCH: 7 LOSS:  3555.033447265625\n",
      "IMAGE EPOCH: 8 LOSS:  3418.604248046875\n",
      "IMAGE EPOCH: 9 LOSS:  3996.39599609375\n",
      "IMAGE EPOCH: 10 LOSS:  3690.64599609375\n",
      "IMAGE EPOCH: 11 LOSS:  3227.39990234375\n",
      "IMAGE EPOCH: 12 LOSS:  3589.6845703125\n",
      "IMAGE EPOCH: 13 LOSS:  3707.1142578125\n",
      "IMAGE EPOCH: 14 LOSS:  3271.468017578125\n",
      "IMAGE EPOCH: 15 LOSS:  3308.596923828125\n",
      "IMAGE EPOCH: 16 LOSS:  3571.0927734375\n",
      "IMAGE EPOCH: 17 LOSS:  3388.1181640625\n",
      "IMAGE EPOCH: 18 LOSS:  3198.577392578125\n",
      "IMAGE EPOCH: 19 LOSS:  3378.3681640625\n",
      "IMAGE EPOCH: 20 LOSS:  3412.977294921875\n",
      "IMAGE EPOCH: 21 LOSS:  3220.901611328125\n",
      "IMAGE EPOCH: 22 LOSS:  3248.7431640625\n",
      "IMAGE EPOCH: 23 LOSS:  3360.16650390625\n",
      "IMAGE EPOCH: 24 LOSS:  3269.1357421875\n",
      "IMAGE EPOCH: 25 LOSS:  3190.381591796875\n",
      "IMAGE EPOCH: 26 LOSS:  3276.3330078125\n",
      "IMAGE EPOCH: 27 LOSS:  3286.579833984375\n",
      "IMAGE EPOCH: 28 LOSS:  3193.6494140625\n",
      "IMAGE EPOCH: 29 LOSS:  3206.6572265625\n",
      "IMAGE EPOCH: 30 LOSS:  3263.3857421875\n",
      "IMAGE EPOCH: 31 LOSS:  3220.953369140625\n",
      "IMAGE EPOCH: 32 LOSS:  3177.48681640625\n",
      "IMAGE EPOCH: 33 LOSS:  3216.499267578125\n",
      "IMAGE EPOCH: 34 LOSS:  3227.77978515625\n",
      "IMAGE EPOCH: 35 LOSS:  3183.74072265625\n",
      "IMAGE EPOCH: 36 LOSS:  3183.4443359375\n",
      "IMAGE EPOCH: 37 LOSS:  3210.99951171875\n",
      "IMAGE EPOCH: 38 LOSS:  3194.80908203125\n",
      "IMAGE EPOCH: 39 LOSS:  3172.9814453125\n",
      "IMAGE EPOCH: 40 LOSS:  3189.267333984375\n",
      "IMAGE EPOCH: 41 LOSS:  3195.6767578125\n",
      "IMAGE EPOCH: 42 LOSS:  3175.3701171875\n",
      "IMAGE EPOCH: 43 LOSS:  3174.46044921875\n",
      "IMAGE EPOCH: 44 LOSS:  3187.877685546875\n",
      "IMAGE EPOCH: 45 LOSS:  3180.15283203125\n",
      "IMAGE EPOCH: 46 LOSS:  3168.74365234375\n",
      "IMAGE EPOCH: 47 LOSS:  3176.52783203125\n",
      "IMAGE EPOCH: 48 LOSS:  3180.2734375\n",
      "IMAGE EPOCH: 49 LOSS:  3169.937255859375\n",
      "IMAGE EPOCH: 50 LOSS:  3168.506103515625\n",
      "IMAGE EPOCH: 51 LOSS:  3175.160400390625\n",
      "IMAGE EPOCH: 52 LOSS:  3171.812255859375\n",
      "IMAGE EPOCH: 53 LOSS:  3165.9052734375\n",
      "IMAGE EPOCH: 54 LOSS:  3169.10107421875\n",
      "IMAGE EPOCH: 55 LOSS:  3170.76708984375\n",
      "IMAGE EPOCH: 56 LOSS:  3165.8623046875\n",
      "IMAGE EPOCH: 57 LOSS:  3165.107666015625\n",
      "IMAGE EPOCH: 58 LOSS:  3167.88720703125\n",
      "IMAGE EPOCH: 59 LOSS:  3165.81689453125\n",
      "IMAGE EPOCH: 60 LOSS:  3163.05029296875\n",
      "IMAGE EPOCH: 61 LOSS:  3164.69970703125\n",
      "IMAGE EPOCH: 62 LOSS:  3165.025390625\n",
      "IMAGE EPOCH: 63 LOSS:  3162.294921875\n",
      "IMAGE EPOCH: 64 LOSS:  3162.067138671875\n",
      "IMAGE EPOCH: 65 LOSS:  3163.31884765625\n",
      "IMAGE EPOCH: 66 LOSS:  3161.928466796875\n",
      "IMAGE EPOCH: 67 LOSS:  3160.49560546875\n",
      "IMAGE EPOCH: 68 LOSS:  3161.24560546875\n",
      "IMAGE EPOCH: 69 LOSS:  3161.1044921875\n",
      "IMAGE EPOCH: 70 LOSS:  3159.66015625\n",
      "IMAGE EPOCH: 71 LOSS:  3159.55224609375\n",
      "IMAGE EPOCH: 72 LOSS:  3159.874755859375\n",
      "IMAGE EPOCH: 73 LOSS:  3158.950927734375\n",
      "IMAGE EPOCH: 74 LOSS:  3158.31787109375\n",
      "IMAGE EPOCH: 75 LOSS:  3158.579833984375\n",
      "IMAGE EPOCH: 76 LOSS:  3158.16748046875\n",
      "IMAGE EPOCH: 77 LOSS:  3157.36572265625\n",
      "IMAGE EPOCH: 78 LOSS:  3157.336181640625\n",
      "IMAGE EPOCH: 79 LOSS:  3157.26806640625\n",
      "IMAGE EPOCH: 80 LOSS:  3156.5869140625\n",
      "IMAGE EPOCH: 81 LOSS:  3156.242431640625\n",
      "IMAGE EPOCH: 82 LOSS:  3156.25\n",
      "IMAGE EPOCH: 83 LOSS:  3155.83056640625\n",
      "IMAGE EPOCH: 84 LOSS:  3155.34765625\n",
      "IMAGE EPOCH: 85 LOSS:  3155.24658203125\n",
      "IMAGE EPOCH: 86 LOSS:  3155.01513671875\n",
      "IMAGE EPOCH: 87 LOSS:  3154.55810546875\n",
      "IMAGE EPOCH: 88 LOSS:  3154.334228515625\n",
      "IMAGE EPOCH: 89 LOSS:  3154.173583984375\n",
      "IMAGE EPOCH: 90 LOSS:  3153.796875\n",
      "IMAGE EPOCH: 91 LOSS:  3153.49658203125\n",
      "IMAGE EPOCH: 92 LOSS:  3153.341552734375\n",
      "IMAGE EPOCH: 93 LOSS:  3153.05126953125\n",
      "IMAGE EPOCH: 94 LOSS:  3152.7197265625\n",
      "IMAGE EPOCH: 95 LOSS:  3152.530517578125\n",
      "IMAGE EPOCH: 96 LOSS:  3152.305908203125\n",
      "IMAGE EPOCH: 97 LOSS:  3151.9892578125\n",
      "IMAGE EPOCH: 98 LOSS:  3151.75439453125\n",
      "IMAGE EPOCH: 99 LOSS:  3151.55419921875\n",
      "BATCH: 4\n",
      "::: PREPROCESSING: 480\n",
      "::: PREPROCESSING: 490\n",
      "::: PREPROCESSING: 500\n",
      "::: PREPROCESSING: 510\n",
      "::: PREPROCESSING: 520\n",
      "::: PREPROCESSING: 530\n",
      "::: PREPROCESSING: 540\n",
      "::: PREPROCESSING: 550\n",
      "::: PREPROCESSING: 560\n",
      "::: PREPROCESSING: 570\n",
      "::: PREPROCESSING: 580\n",
      "::: PREPROCESSING: 590\n",
      "IMAGE EPOCH: 0 LOSS:  9405.00390625\n",
      "IMAGE EPOCH: 1 LOSS:  8779.4765625\n",
      "IMAGE EPOCH: 2 LOSS:  8491.314453125\n",
      "IMAGE EPOCH: 3 LOSS:  8849.1650390625\n",
      "IMAGE EPOCH: 4 LOSS:  9066.958984375\n",
      "IMAGE EPOCH: 5 LOSS:  8721.5498046875\n",
      "IMAGE EPOCH: 6 LOSS:  8393.8046875\n",
      "IMAGE EPOCH: 7 LOSS:  8492.802734375\n",
      "IMAGE EPOCH: 8 LOSS:  8700.396484375\n",
      "IMAGE EPOCH: 9 LOSS:  8634.728515625\n",
      "IMAGE EPOCH: 10 LOSS:  8385.4853515625\n",
      "IMAGE EPOCH: 11 LOSS:  8287.2099609375\n",
      "IMAGE EPOCH: 12 LOSS:  8407.75\n",
      "IMAGE EPOCH: 13 LOSS:  8482.65625\n",
      "IMAGE EPOCH: 14 LOSS:  8366.435546875\n",
      "IMAGE EPOCH: 15 LOSS:  8233.078125\n",
      "IMAGE EPOCH: 16 LOSS:  8246.515625\n",
      "IMAGE EPOCH: 17 LOSS:  8328.4560546875\n",
      "IMAGE EPOCH: 18 LOSS:  8326.6318359375\n",
      "IMAGE EPOCH: 19 LOSS:  8238.0859375\n",
      "IMAGE EPOCH: 20 LOSS:  8182.916015625\n",
      "IMAGE EPOCH: 21 LOSS:  8216.4228515625\n",
      "IMAGE EPOCH: 22 LOSS:  8259.078125\n",
      "IMAGE EPOCH: 23 LOSS:  8233.205078125\n",
      "IMAGE EPOCH: 24 LOSS:  8177.07421875\n",
      "IMAGE EPOCH: 25 LOSS:  8165.83984375\n",
      "IMAGE EPOCH: 26 LOSS:  8195.419921875\n",
      "IMAGE EPOCH: 27 LOSS:  8205.5986328125\n",
      "IMAGE EPOCH: 28 LOSS:  8176.013671875\n",
      "IMAGE EPOCH: 29 LOSS:  8146.60595703125\n",
      "IMAGE EPOCH: 30 LOSS:  8150.5244140625\n",
      "IMAGE EPOCH: 31 LOSS:  8167.75830078125\n",
      "IMAGE EPOCH: 32 LOSS:  8162.98046875\n",
      "IMAGE EPOCH: 33 LOSS:  8139.85546875\n",
      "IMAGE EPOCH: 34 LOSS:  8128.49951171875\n",
      "IMAGE EPOCH: 35 LOSS:  8136.4326171875\n",
      "IMAGE EPOCH: 36 LOSS:  8142.1142578125\n",
      "IMAGE EPOCH: 37 LOSS:  8131.41162109375\n",
      "IMAGE EPOCH: 38 LOSS:  8116.490234375\n",
      "IMAGE EPOCH: 39 LOSS:  8113.58837890625\n",
      "IMAGE EPOCH: 40 LOSS:  8118.677734375\n",
      "IMAGE EPOCH: 41 LOSS:  8116.9189453125\n",
      "IMAGE EPOCH: 42 LOSS:  8106.65234375\n",
      "IMAGE EPOCH: 43 LOSS:  8099.43359375\n",
      "IMAGE EPOCH: 44 LOSS:  8100.43505859375\n",
      "IMAGE EPOCH: 45 LOSS:  8102.0068359375\n",
      "IMAGE EPOCH: 46 LOSS:  8097.2431640625\n",
      "IMAGE EPOCH: 47 LOSS:  8089.8916015625\n",
      "IMAGE EPOCH: 48 LOSS:  8086.88134765625\n",
      "IMAGE EPOCH: 49 LOSS:  8087.51416015625\n",
      "IMAGE EPOCH: 50 LOSS:  8085.85791015625\n",
      "IMAGE EPOCH: 51 LOSS:  8080.6142578125\n",
      "IMAGE EPOCH: 52 LOSS:  8076.3046875\n",
      "IMAGE EPOCH: 53 LOSS:  8075.38134765625\n",
      "IMAGE EPOCH: 54 LOSS:  8074.865234375\n",
      "IMAGE EPOCH: 55 LOSS:  8071.798828125\n",
      "IMAGE EPOCH: 56 LOSS:  8067.6611328125\n",
      "IMAGE EPOCH: 57 LOSS:  8065.255859375\n",
      "IMAGE EPOCH: 58 LOSS:  8064.1611328125\n",
      "IMAGE EPOCH: 59 LOSS:  8061.99072265625\n",
      "IMAGE EPOCH: 60 LOSS:  8058.47509765625\n",
      "IMAGE EPOCH: 61 LOSS:  8055.4912109375\n",
      "IMAGE EPOCH: 62 LOSS:  8053.7021484375\n",
      "IMAGE EPOCH: 63 LOSS:  8051.69873046875\n",
      "IMAGE EPOCH: 64 LOSS:  8048.5302734375\n",
      "IMAGE EPOCH: 65 LOSS:  8045.1708984375\n",
      "IMAGE EPOCH: 66 LOSS:  8042.7392578125\n",
      "IMAGE EPOCH: 67 LOSS:  8040.78564453125\n",
      "IMAGE EPOCH: 68 LOSS:  8038.30859375\n",
      "IMAGE EPOCH: 69 LOSS:  8035.4658203125\n",
      "IMAGE EPOCH: 70 LOSS:  8033.0478515625\n",
      "IMAGE EPOCH: 71 LOSS:  8031.01025390625\n",
      "IMAGE EPOCH: 72 LOSS:  8028.6220703125\n",
      "IMAGE EPOCH: 73 LOSS:  8025.787109375\n",
      "IMAGE EPOCH: 74 LOSS:  8023.12548828125\n",
      "IMAGE EPOCH: 75 LOSS:  8020.8388671875\n",
      "IMAGE EPOCH: 76 LOSS:  8018.49169921875\n",
      "IMAGE EPOCH: 77 LOSS:  8015.908203125\n",
      "IMAGE EPOCH: 78 LOSS:  8013.45361328125\n",
      "IMAGE EPOCH: 79 LOSS:  8011.2890625\n",
      "IMAGE EPOCH: 80 LOSS:  8009.07470703125\n",
      "IMAGE EPOCH: 81 LOSS:  8006.60205078125\n",
      "IMAGE EPOCH: 82 LOSS:  8004.109375\n",
      "IMAGE EPOCH: 83 LOSS:  8001.7861328125\n",
      "IMAGE EPOCH: 84 LOSS:  7999.45361328125\n",
      "IMAGE EPOCH: 85 LOSS:  7996.974609375\n",
      "IMAGE EPOCH: 86 LOSS:  7994.5078125\n",
      "IMAGE EPOCH: 87 LOSS:  7992.19384765625\n",
      "IMAGE EPOCH: 88 LOSS:  7989.89990234375\n",
      "IMAGE EPOCH: 89 LOSS:  7987.48876953125\n",
      "IMAGE EPOCH: 90 LOSS:  7985.046875\n",
      "IMAGE EPOCH: 91 LOSS:  7982.6611328125\n",
      "IMAGE EPOCH: 92 LOSS:  7980.259765625\n",
      "IMAGE EPOCH: 93 LOSS:  7977.787109375\n",
      "IMAGE EPOCH: 94 LOSS:  7975.32421875\n",
      "IMAGE EPOCH: 95 LOSS:  7972.91748046875\n",
      "IMAGE EPOCH: 96 LOSS:  7970.5009765625\n",
      "IMAGE EPOCH: 97 LOSS:  7968.0244140625\n",
      "IMAGE EPOCH: 98 LOSS:  7965.5400390625\n",
      "IMAGE EPOCH: 99 LOSS:  7963.07763671875\n",
      "EPOCH: 2 SHARPNESS: 3.0 FG VOXEL SIZE: 64\n",
      "BATCH: 0\n",
      "::: PREPROCESSING: 0\n",
      "::: PREPROCESSING: 10\n",
      "::: PREPROCESSING: 20\n",
      "::: PREPROCESSING: 30\n",
      "::: PREPROCESSING: 40\n",
      "::: PREPROCESSING: 50\n",
      "IMAGE EPOCH: 0 LOSS:  5775.390625\n",
      "IMAGE EPOCH: 1 LOSS:  5677.38623046875\n",
      "IMAGE EPOCH: 2 LOSS:  5518.41259765625\n",
      "IMAGE EPOCH: 3 LOSS:  5347.55712890625\n",
      "IMAGE EPOCH: 4 LOSS:  5198.30712890625\n",
      "IMAGE EPOCH: 5 LOSS:  5083.89111328125\n",
      "IMAGE EPOCH: 6 LOSS:  4999.8330078125\n",
      "IMAGE EPOCH: 7 LOSS:  4929.05517578125\n",
      "IMAGE EPOCH: 8 LOSS:  4857.05078125\n",
      "IMAGE EPOCH: 9 LOSS:  4772.859375\n",
      "IMAGE EPOCH: 10 LOSS:  4680.48974609375\n",
      "IMAGE EPOCH: 11 LOSS:  4597.59716796875\n",
      "IMAGE EPOCH: 12 LOSS:  4540.3359375\n",
      "IMAGE EPOCH: 13 LOSS:  4509.07958984375\n",
      "IMAGE EPOCH: 14 LOSS:  4486.40771484375\n",
      "IMAGE EPOCH: 15 LOSS:  4454.1640625\n",
      "IMAGE EPOCH: 16 LOSS:  4407.5888671875\n",
      "IMAGE EPOCH: 17 LOSS:  4354.08447265625\n",
      "IMAGE EPOCH: 18 LOSS:  4304.056640625\n",
      "IMAGE EPOCH: 19 LOSS:  4264.169921875\n",
      "IMAGE EPOCH: 20 LOSS:  4232.94140625\n",
      "IMAGE EPOCH: 21 LOSS:  4201.10302734375\n",
      "IMAGE EPOCH: 22 LOSS:  4159.1123046875\n",
      "IMAGE EPOCH: 23 LOSS:  4109.50244140625\n",
      "IMAGE EPOCH: 24 LOSS:  4066.471923828125\n",
      "IMAGE EPOCH: 25 LOSS:  4045.09716796875\n",
      "IMAGE EPOCH: 26 LOSS:  4019.870361328125\n",
      "IMAGE EPOCH: 27 LOSS:  4000.217041015625\n",
      "IMAGE EPOCH: 28 LOSS:  3973.4501953125\n",
      "IMAGE EPOCH: 29 LOSS:  3930.8974609375\n",
      "IMAGE EPOCH: 30 LOSS:  3889.03125\n",
      "IMAGE EPOCH: 31 LOSS:  3860.973876953125\n",
      "IMAGE EPOCH: 32 LOSS:  3825.48046875\n",
      "IMAGE EPOCH: 33 LOSS:  3793.135986328125\n",
      "IMAGE EPOCH: 34 LOSS:  3759.138916015625\n",
      "IMAGE EPOCH: 35 LOSS:  3720.62158203125\n",
      "IMAGE EPOCH: 36 LOSS:  3684.149658203125\n",
      "IMAGE EPOCH: 37 LOSS:  3654.61767578125\n",
      "IMAGE EPOCH: 38 LOSS:  3628.3466796875\n",
      "IMAGE EPOCH: 39 LOSS:  3603.6689453125\n",
      "IMAGE EPOCH: 40 LOSS:  3584.392578125\n",
      "IMAGE EPOCH: 41 LOSS:  3565.8447265625\n",
      "IMAGE EPOCH: 42 LOSS:  3546.973388671875\n",
      "IMAGE EPOCH: 43 LOSS:  3529.90869140625\n",
      "IMAGE EPOCH: 44 LOSS:  3512.3603515625\n",
      "IMAGE EPOCH: 45 LOSS:  3497.554931640625\n",
      "IMAGE EPOCH: 46 LOSS:  3483.6845703125\n",
      "IMAGE EPOCH: 47 LOSS:  3467.716552734375\n",
      "IMAGE EPOCH: 48 LOSS:  3450.541259765625\n",
      "IMAGE EPOCH: 49 LOSS:  3432.773193359375\n",
      "IMAGE EPOCH: 50 LOSS:  3416.57568359375\n",
      "IMAGE EPOCH: 51 LOSS:  3403.55810546875\n",
      "IMAGE EPOCH: 52 LOSS:  3391.68310546875\n",
      "IMAGE EPOCH: 53 LOSS:  3381.117919921875\n",
      "IMAGE EPOCH: 54 LOSS:  3370.861572265625\n",
      "IMAGE EPOCH: 55 LOSS:  3361.6943359375\n",
      "IMAGE EPOCH: 56 LOSS:  3353.40234375\n",
      "IMAGE EPOCH: 57 LOSS:  3346.02099609375\n",
      "IMAGE EPOCH: 58 LOSS:  3339.48779296875\n",
      "IMAGE EPOCH: 59 LOSS:  3333.280029296875\n",
      "IMAGE EPOCH: 60 LOSS:  3327.266357421875\n",
      "IMAGE EPOCH: 61 LOSS:  3321.183837890625\n",
      "IMAGE EPOCH: 62 LOSS:  3315.715576171875\n",
      "IMAGE EPOCH: 63 LOSS:  3310.839599609375\n",
      "IMAGE EPOCH: 64 LOSS:  3306.6689453125\n",
      "IMAGE EPOCH: 65 LOSS:  3302.6171875\n",
      "IMAGE EPOCH: 66 LOSS:  3298.65869140625\n",
      "IMAGE EPOCH: 67 LOSS:  3294.821044921875\n",
      "IMAGE EPOCH: 68 LOSS:  3291.113037109375\n",
      "IMAGE EPOCH: 69 LOSS:  3287.73388671875\n",
      "IMAGE EPOCH: 70 LOSS:  3284.3984375\n",
      "IMAGE EPOCH: 71 LOSS:  3281.37841796875\n",
      "IMAGE EPOCH: 72 LOSS:  3278.31787109375\n",
      "IMAGE EPOCH: 73 LOSS:  3275.32763671875\n",
      "IMAGE EPOCH: 74 LOSS:  3272.43505859375\n",
      "IMAGE EPOCH: 75 LOSS:  3269.68310546875\n",
      "IMAGE EPOCH: 76 LOSS:  3267.271240234375\n",
      "IMAGE EPOCH: 77 LOSS:  3264.8896484375\n",
      "IMAGE EPOCH: 78 LOSS:  3262.43212890625\n",
      "IMAGE EPOCH: 79 LOSS:  3260.0673828125\n",
      "IMAGE EPOCH: 80 LOSS:  3257.711181640625\n",
      "IMAGE EPOCH: 81 LOSS:  3255.4052734375\n",
      "IMAGE EPOCH: 82 LOSS:  3253.2900390625\n",
      "IMAGE EPOCH: 83 LOSS:  3251.2255859375\n",
      "IMAGE EPOCH: 84 LOSS:  3249.18408203125\n",
      "IMAGE EPOCH: 85 LOSS:  3247.20703125\n",
      "IMAGE EPOCH: 86 LOSS:  3245.223388671875\n",
      "IMAGE EPOCH: 87 LOSS:  3243.29541015625\n",
      "IMAGE EPOCH: 88 LOSS:  3241.4921875\n",
      "IMAGE EPOCH: 89 LOSS:  3239.70654296875\n",
      "IMAGE EPOCH: 90 LOSS:  3237.925048828125\n",
      "IMAGE EPOCH: 91 LOSS:  3236.212158203125\n",
      "IMAGE EPOCH: 92 LOSS:  3234.537109375\n",
      "IMAGE EPOCH: 93 LOSS:  3232.87255859375\n",
      "IMAGE EPOCH: 94 LOSS:  3231.248046875\n",
      "IMAGE EPOCH: 95 LOSS:  3229.653564453125\n",
      "IMAGE EPOCH: 96 LOSS:  3228.046630859375\n",
      "IMAGE EPOCH: 97 LOSS:  3226.434814453125\n",
      "IMAGE EPOCH: 98 LOSS:  3224.8388671875\n",
      "IMAGE EPOCH: 99 LOSS:  3223.2568359375\n",
      "BATCH: 1\n",
      "::: PREPROCESSING: 60\n",
      "::: PREPROCESSING: 70\n",
      "::: PREPROCESSING: 80\n",
      "::: PREPROCESSING: 90\n",
      "::: PREPROCESSING: 100\n",
      "::: PREPROCESSING: 110\n",
      "IMAGE EPOCH: 0 LOSS:  4643.8642578125\n",
      "IMAGE EPOCH: 1 LOSS:  3620.431640625\n",
      "IMAGE EPOCH: 2 LOSS:  3397.06640625\n",
      "IMAGE EPOCH: 3 LOSS:  3036.4150390625\n",
      "IMAGE EPOCH: 4 LOSS:  3503.732666015625\n",
      "IMAGE EPOCH: 5 LOSS:  3211.6982421875\n",
      "IMAGE EPOCH: 6 LOSS:  3520.560791015625\n",
      "IMAGE EPOCH: 7 LOSS:  3229.6943359375\n",
      "IMAGE EPOCH: 8 LOSS:  3312.27197265625\n",
      "IMAGE EPOCH: 9 LOSS:  3080.509033203125\n",
      "IMAGE EPOCH: 10 LOSS:  2998.425537109375\n",
      "IMAGE EPOCH: 11 LOSS:  3012.814208984375\n",
      "IMAGE EPOCH: 12 LOSS:  2909.865478515625\n",
      "IMAGE EPOCH: 13 LOSS:  3044.695556640625\n",
      "IMAGE EPOCH: 14 LOSS:  2943.771240234375\n",
      "IMAGE EPOCH: 15 LOSS:  3035.8154296875\n",
      "IMAGE EPOCH: 16 LOSS:  3014.7666015625\n",
      "IMAGE EPOCH: 17 LOSS:  2969.936767578125\n",
      "IMAGE EPOCH: 18 LOSS:  2988.798828125\n",
      "IMAGE EPOCH: 19 LOSS:  2887.396240234375\n",
      "IMAGE EPOCH: 20 LOSS:  2910.508544921875\n",
      "IMAGE EPOCH: 21 LOSS:  2876.146240234375\n",
      "IMAGE EPOCH: 22 LOSS:  2855.97119140625\n",
      "IMAGE EPOCH: 23 LOSS:  2888.507080078125\n",
      "IMAGE EPOCH: 24 LOSS:  2858.293701171875\n",
      "IMAGE EPOCH: 25 LOSS:  2891.68310546875\n",
      "IMAGE EPOCH: 26 LOSS:  2883.87939453125\n",
      "IMAGE EPOCH: 27 LOSS:  2864.15283203125\n",
      "IMAGE EPOCH: 28 LOSS:  2875.194580078125\n",
      "IMAGE EPOCH: 29 LOSS:  2843.652587890625\n",
      "IMAGE EPOCH: 30 LOSS:  2843.7099609375\n",
      "IMAGE EPOCH: 31 LOSS:  2837.19140625\n",
      "IMAGE EPOCH: 32 LOSS:  2822.037841796875\n",
      "IMAGE EPOCH: 33 LOSS:  2836.779541015625\n",
      "IMAGE EPOCH: 34 LOSS:  2830.2373046875\n",
      "IMAGE EPOCH: 35 LOSS:  2832.930908203125\n",
      "IMAGE EPOCH: 36 LOSS:  2838.24267578125\n",
      "IMAGE EPOCH: 37 LOSS:  2827.331298828125\n",
      "IMAGE EPOCH: 38 LOSS:  2831.163818359375\n",
      "IMAGE EPOCH: 39 LOSS:  2824.83056640625\n",
      "IMAGE EPOCH: 40 LOSS:  2817.4658203125\n",
      "IMAGE EPOCH: 41 LOSS:  2819.58154296875\n",
      "IMAGE EPOCH: 42 LOSS:  2813.21630859375\n",
      "IMAGE EPOCH: 43 LOSS:  2815.587890625\n",
      "IMAGE EPOCH: 44 LOSS:  2816.74560546875\n",
      "IMAGE EPOCH: 45 LOSS:  2813.505859375\n",
      "IMAGE EPOCH: 46 LOSS:  2816.673583984375\n",
      "IMAGE EPOCH: 47 LOSS:  2813.580078125\n",
      "IMAGE EPOCH: 48 LOSS:  2812.11376953125\n",
      "IMAGE EPOCH: 49 LOSS:  2811.533203125\n",
      "IMAGE EPOCH: 50 LOSS:  2807.21630859375\n",
      "IMAGE EPOCH: 51 LOSS:  2807.64794921875\n",
      "IMAGE EPOCH: 52 LOSS:  2806.0771484375\n",
      "IMAGE EPOCH: 53 LOSS:  2804.96826171875\n",
      "IMAGE EPOCH: 54 LOSS:  2805.9267578125\n",
      "IMAGE EPOCH: 55 LOSS:  2804.2939453125\n",
      "IMAGE EPOCH: 56 LOSS:  2804.795166015625\n",
      "IMAGE EPOCH: 57 LOSS:  2804.085693359375\n",
      "IMAGE EPOCH: 58 LOSS:  2802.506103515625\n",
      "IMAGE EPOCH: 59 LOSS:  2802.367919921875\n",
      "IMAGE EPOCH: 60 LOSS:  2800.66455078125\n",
      "IMAGE EPOCH: 61 LOSS:  2800.1884765625\n",
      "IMAGE EPOCH: 62 LOSS:  2799.635986328125\n",
      "IMAGE EPOCH: 63 LOSS:  2798.615478515625\n",
      "IMAGE EPOCH: 64 LOSS:  2798.7578125\n",
      "IMAGE EPOCH: 65 LOSS:  2797.9658203125\n",
      "IMAGE EPOCH: 66 LOSS:  2797.598388671875\n",
      "IMAGE EPOCH: 67 LOSS:  2797.15576171875\n",
      "IMAGE EPOCH: 68 LOSS:  2796.174072265625\n",
      "IMAGE EPOCH: 69 LOSS:  2795.827392578125\n",
      "IMAGE EPOCH: 70 LOSS:  2794.938232421875\n",
      "IMAGE EPOCH: 71 LOSS:  2794.32958984375\n",
      "IMAGE EPOCH: 72 LOSS:  2793.875\n",
      "IMAGE EPOCH: 73 LOSS:  2793.187744140625\n",
      "IMAGE EPOCH: 74 LOSS:  2792.9345703125\n",
      "IMAGE EPOCH: 75 LOSS:  2792.36083984375\n",
      "IMAGE EPOCH: 76 LOSS:  2791.89599609375\n",
      "IMAGE EPOCH: 77 LOSS:  2791.47607421875\n",
      "IMAGE EPOCH: 78 LOSS:  2790.84130859375\n",
      "IMAGE EPOCH: 79 LOSS:  2790.424560546875\n",
      "IMAGE EPOCH: 80 LOSS:  2789.8232421875\n",
      "IMAGE EPOCH: 81 LOSS:  2789.328857421875\n",
      "IMAGE EPOCH: 82 LOSS:  2788.901611328125\n",
      "IMAGE EPOCH: 83 LOSS:  2788.380859375\n",
      "IMAGE EPOCH: 84 LOSS:  2788.005615234375\n",
      "IMAGE EPOCH: 85 LOSS:  2787.515869140625\n",
      "IMAGE EPOCH: 86 LOSS:  2787.079833984375\n",
      "IMAGE EPOCH: 87 LOSS:  2786.64501953125\n",
      "IMAGE EPOCH: 88 LOSS:  2786.13427734375\n",
      "IMAGE EPOCH: 89 LOSS:  2785.701904296875\n",
      "IMAGE EPOCH: 90 LOSS:  2785.204345703125\n",
      "IMAGE EPOCH: 91 LOSS:  2784.76025390625\n",
      "IMAGE EPOCH: 92 LOSS:  2784.321044921875\n",
      "IMAGE EPOCH: 93 LOSS:  2783.85986328125\n",
      "IMAGE EPOCH: 94 LOSS:  2783.45263671875\n",
      "IMAGE EPOCH: 95 LOSS:  2783.001953125\n",
      "IMAGE EPOCH: 96 LOSS:  2782.582275390625\n",
      "IMAGE EPOCH: 97 LOSS:  2782.14599609375\n",
      "IMAGE EPOCH: 98 LOSS:  2781.697265625\n",
      "IMAGE EPOCH: 99 LOSS:  2781.273193359375\n",
      "BATCH: 2\n",
      "::: PREPROCESSING: 120\n",
      "::: PREPROCESSING: 130\n",
      "::: PREPROCESSING: 140\n",
      "::: PREPROCESSING: 150\n",
      "::: PREPROCESSING: 160\n",
      "::: PREPROCESSING: 170\n",
      "IMAGE EPOCH: 0 LOSS:  6136.85400390625\n",
      "IMAGE EPOCH: 1 LOSS:  3762.095458984375\n",
      "IMAGE EPOCH: 2 LOSS:  3873.208984375\n",
      "IMAGE EPOCH: 3 LOSS:  5173.90478515625\n",
      "IMAGE EPOCH: 4 LOSS:  4511.89501953125\n",
      "IMAGE EPOCH: 5 LOSS:  3253.650390625\n",
      "IMAGE EPOCH: 6 LOSS:  3365.19384765625\n",
      "IMAGE EPOCH: 7 LOSS:  4130.609375\n",
      "IMAGE EPOCH: 8 LOSS:  4046.509765625\n",
      "IMAGE EPOCH: 9 LOSS:  3358.231689453125\n",
      "IMAGE EPOCH: 10 LOSS:  3109.45947265625\n",
      "IMAGE EPOCH: 11 LOSS:  3432.718505859375\n",
      "IMAGE EPOCH: 12 LOSS:  3659.513427734375\n",
      "IMAGE EPOCH: 13 LOSS:  3441.531494140625\n",
      "IMAGE EPOCH: 14 LOSS:  3084.08203125\n",
      "IMAGE EPOCH: 15 LOSS:  2983.711181640625\n",
      "IMAGE EPOCH: 16 LOSS:  3140.030029296875\n",
      "IMAGE EPOCH: 17 LOSS:  3257.928955078125\n",
      "IMAGE EPOCH: 18 LOSS:  3161.24462890625\n",
      "IMAGE EPOCH: 19 LOSS:  2970.62255859375\n",
      "IMAGE EPOCH: 20 LOSS:  2903.1513671875\n",
      "IMAGE EPOCH: 21 LOSS:  2996.864013671875\n",
      "IMAGE EPOCH: 22 LOSS:  3094.304443359375\n",
      "IMAGE EPOCH: 23 LOSS:  3068.18408203125\n",
      "IMAGE EPOCH: 24 LOSS:  2964.3876953125\n",
      "IMAGE EPOCH: 25 LOSS:  2903.93212890625\n",
      "IMAGE EPOCH: 26 LOSS:  2928.085693359375\n",
      "IMAGE EPOCH: 27 LOSS:  2978.908203125\n",
      "IMAGE EPOCH: 28 LOSS:  2986.84033203125\n",
      "IMAGE EPOCH: 29 LOSS:  2941.03515625\n",
      "IMAGE EPOCH: 30 LOSS:  2887.185546875\n",
      "IMAGE EPOCH: 31 LOSS:  2873.056396484375\n",
      "IMAGE EPOCH: 32 LOSS:  2898.234375\n",
      "IMAGE EPOCH: 33 LOSS:  2920.896484375\n",
      "IMAGE EPOCH: 34 LOSS:  2910.6435546875\n",
      "IMAGE EPOCH: 35 LOSS:  2879.527587890625\n",
      "IMAGE EPOCH: 36 LOSS:  2858.999755859375\n",
      "IMAGE EPOCH: 37 LOSS:  2862.18310546875\n",
      "IMAGE EPOCH: 38 LOSS:  2876.193115234375\n",
      "IMAGE EPOCH: 39 LOSS:  2881.36083984375\n",
      "IMAGE EPOCH: 40 LOSS:  2870.404541015625\n",
      "IMAGE EPOCH: 41 LOSS:  2852.458251953125\n",
      "IMAGE EPOCH: 42 LOSS:  2842.28076171875\n",
      "IMAGE EPOCH: 43 LOSS:  2845.024169921875\n",
      "IMAGE EPOCH: 44 LOSS:  2851.87744140625\n",
      "IMAGE EPOCH: 45 LOSS:  2851.4248046875\n",
      "IMAGE EPOCH: 46 LOSS:  2842.437255859375\n",
      "IMAGE EPOCH: 47 LOSS:  2833.046875\n",
      "IMAGE EPOCH: 48 LOSS:  2830.150146484375\n",
      "IMAGE EPOCH: 49 LOSS:  2832.96875\n",
      "IMAGE EPOCH: 50 LOSS:  2835.75439453125\n",
      "IMAGE EPOCH: 51 LOSS:  2834.08544921875\n",
      "IMAGE EPOCH: 52 LOSS:  2828.59033203125\n",
      "IMAGE EPOCH: 53 LOSS:  2823.654052734375\n",
      "IMAGE EPOCH: 54 LOSS:  2822.5576171875\n",
      "IMAGE EPOCH: 55 LOSS:  2824.087158203125\n",
      "IMAGE EPOCH: 56 LOSS:  2824.43505859375\n",
      "IMAGE EPOCH: 57 LOSS:  2821.78076171875\n",
      "IMAGE EPOCH: 58 LOSS:  2817.836181640625\n",
      "IMAGE EPOCH: 59 LOSS:  2815.240478515625\n",
      "IMAGE EPOCH: 60 LOSS:  2814.7265625\n",
      "IMAGE EPOCH: 61 LOSS:  2814.93505859375\n",
      "IMAGE EPOCH: 62 LOSS:  2814.163330078125\n",
      "IMAGE EPOCH: 63 LOSS:  2811.98193359375\n",
      "IMAGE EPOCH: 64 LOSS:  2809.474853515625\n",
      "IMAGE EPOCH: 65 LOSS:  2807.9677734375\n",
      "IMAGE EPOCH: 66 LOSS:  2807.58203125\n",
      "IMAGE EPOCH: 67 LOSS:  2807.25048828125\n",
      "IMAGE EPOCH: 68 LOSS:  2806.07470703125\n",
      "IMAGE EPOCH: 69 LOSS:  2804.2763671875\n",
      "IMAGE EPOCH: 70 LOSS:  2802.69775390625\n",
      "IMAGE EPOCH: 71 LOSS:  2801.81201171875\n",
      "IMAGE EPOCH: 72 LOSS:  2801.35693359375\n",
      "IMAGE EPOCH: 73 LOSS:  2800.751953125\n",
      "IMAGE EPOCH: 74 LOSS:  2799.69384765625\n",
      "IMAGE EPOCH: 75 LOSS:  2798.41845703125\n",
      "IMAGE EPOCH: 76 LOSS:  2797.394775390625\n",
      "IMAGE EPOCH: 77 LOSS:  2796.783447265625\n",
      "IMAGE EPOCH: 78 LOSS:  2796.294677734375\n",
      "IMAGE EPOCH: 79 LOSS:  2795.579833984375\n",
      "IMAGE EPOCH: 80 LOSS:  2794.625\n",
      "IMAGE EPOCH: 81 LOSS:  2793.685791015625\n",
      "IMAGE EPOCH: 82 LOSS:  2792.947509765625\n",
      "IMAGE EPOCH: 83 LOSS:  2792.36767578125\n",
      "IMAGE EPOCH: 84 LOSS:  2791.765380859375\n",
      "IMAGE EPOCH: 85 LOSS:  2791.01904296875\n",
      "IMAGE EPOCH: 86 LOSS:  2790.18505859375\n",
      "IMAGE EPOCH: 87 LOSS:  2789.41845703125\n",
      "IMAGE EPOCH: 88 LOSS:  2788.78759765625\n",
      "IMAGE EPOCH: 89 LOSS:  2788.21240234375\n",
      "IMAGE EPOCH: 90 LOSS:  2787.57763671875\n",
      "IMAGE EPOCH: 91 LOSS:  2786.870849609375\n",
      "IMAGE EPOCH: 92 LOSS:  2786.16943359375\n",
      "IMAGE EPOCH: 93 LOSS:  2785.537353515625\n",
      "IMAGE EPOCH: 94 LOSS:  2784.96728515625\n",
      "IMAGE EPOCH: 95 LOSS:  2784.400146484375\n",
      "IMAGE EPOCH: 96 LOSS:  2783.794189453125\n",
      "IMAGE EPOCH: 97 LOSS:  2783.164794921875\n",
      "IMAGE EPOCH: 98 LOSS:  2782.56201171875\n",
      "IMAGE EPOCH: 99 LOSS:  2782.0107421875\n",
      "BATCH: 3\n",
      "::: PREPROCESSING: 180\n",
      "::: PREPROCESSING: 190\n",
      "::: PREPROCESSING: 200\n",
      "::: PREPROCESSING: 210\n",
      "::: PREPROCESSING: 220\n",
      "::: PREPROCESSING: 230\n",
      "IMAGE EPOCH: 0 LOSS:  3407.58935546875\n",
      "IMAGE EPOCH: 1 LOSS:  3171.705810546875\n",
      "IMAGE EPOCH: 2 LOSS:  2983.389892578125\n",
      "IMAGE EPOCH: 3 LOSS:  2995.655517578125\n",
      "IMAGE EPOCH: 4 LOSS:  3093.11767578125\n",
      "IMAGE EPOCH: 5 LOSS:  3094.34912109375\n",
      "IMAGE EPOCH: 6 LOSS:  2970.49853515625\n",
      "IMAGE EPOCH: 7 LOSS:  2844.319580078125\n",
      "IMAGE EPOCH: 8 LOSS:  2822.554443359375\n",
      "IMAGE EPOCH: 9 LOSS:  2875.400146484375\n",
      "IMAGE EPOCH: 10 LOSS:  2892.959716796875\n",
      "IMAGE EPOCH: 11 LOSS:  2829.023193359375\n",
      "IMAGE EPOCH: 12 LOSS:  2745.778564453125\n",
      "IMAGE EPOCH: 13 LOSS:  2720.35693359375\n",
      "IMAGE EPOCH: 14 LOSS:  2751.364990234375\n",
      "IMAGE EPOCH: 15 LOSS:  2774.67919921875\n",
      "IMAGE EPOCH: 16 LOSS:  2749.9150390625\n",
      "IMAGE EPOCH: 17 LOSS:  2703.158447265625\n",
      "IMAGE EPOCH: 18 LOSS:  2683.33935546875\n",
      "IMAGE EPOCH: 19 LOSS:  2698.755859375\n",
      "IMAGE EPOCH: 20 LOSS:  2712.99169921875\n",
      "IMAGE EPOCH: 21 LOSS:  2696.7763671875\n",
      "IMAGE EPOCH: 22 LOSS:  2663.468994140625\n",
      "IMAGE EPOCH: 23 LOSS:  2645.966064453125\n",
      "IMAGE EPOCH: 24 LOSS:  2652.698974609375\n",
      "IMAGE EPOCH: 25 LOSS:  2661.5205078125\n",
      "IMAGE EPOCH: 26 LOSS:  2653.02197265625\n",
      "IMAGE EPOCH: 27 LOSS:  2634.792724609375\n",
      "IMAGE EPOCH: 28 LOSS:  2626.667236328125\n",
      "IMAGE EPOCH: 29 LOSS:  2632.30859375\n",
      "IMAGE EPOCH: 30 LOSS:  2636.723388671875\n",
      "IMAGE EPOCH: 31 LOSS:  2629.251708984375\n",
      "IMAGE EPOCH: 32 LOSS:  2617.3359375\n",
      "IMAGE EPOCH: 33 LOSS:  2613.775146484375\n",
      "IMAGE EPOCH: 34 LOSS:  2618.709228515625\n",
      "IMAGE EPOCH: 35 LOSS:  2621.4423828125\n",
      "IMAGE EPOCH: 36 LOSS:  2616.52392578125\n",
      "IMAGE EPOCH: 37 LOSS:  2609.9521484375\n",
      "IMAGE EPOCH: 38 LOSS:  2608.689697265625\n",
      "IMAGE EPOCH: 39 LOSS:  2610.809326171875\n",
      "IMAGE EPOCH: 40 LOSS:  2609.824462890625\n",
      "IMAGE EPOCH: 41 LOSS:  2604.7109375\n",
      "IMAGE EPOCH: 42 LOSS:  2600.562255859375\n",
      "IMAGE EPOCH: 43 LOSS:  2600.495849609375\n",
      "IMAGE EPOCH: 44 LOSS:  2601.612548828125\n",
      "IMAGE EPOCH: 45 LOSS:  2600.07763671875\n",
      "IMAGE EPOCH: 46 LOSS:  2596.597412109375\n",
      "IMAGE EPOCH: 47 LOSS:  2594.55322265625\n",
      "IMAGE EPOCH: 48 LOSS:  2594.725341796875\n",
      "IMAGE EPOCH: 49 LOSS:  2594.70263671875\n",
      "IMAGE EPOCH: 50 LOSS:  2592.92431640625\n",
      "IMAGE EPOCH: 51 LOSS:  2590.753173828125\n",
      "IMAGE EPOCH: 52 LOSS:  2589.98583984375\n",
      "IMAGE EPOCH: 53 LOSS:  2590.189208984375\n",
      "IMAGE EPOCH: 54 LOSS:  2589.690185546875\n",
      "IMAGE EPOCH: 55 LOSS:  2588.1572265625\n",
      "IMAGE EPOCH: 56 LOSS:  2586.79736328125\n",
      "IMAGE EPOCH: 57 LOSS:  2586.34326171875\n",
      "IMAGE EPOCH: 58 LOSS:  2586.096923828125\n",
      "IMAGE EPOCH: 59 LOSS:  2585.210205078125\n",
      "IMAGE EPOCH: 60 LOSS:  2583.931884765625\n",
      "IMAGE EPOCH: 61 LOSS:  2583.0703125\n",
      "IMAGE EPOCH: 62 LOSS:  2582.737060546875\n",
      "IMAGE EPOCH: 63 LOSS:  2582.323486328125\n",
      "IMAGE EPOCH: 64 LOSS:  2581.50439453125\n",
      "IMAGE EPOCH: 65 LOSS:  2580.640869140625\n",
      "IMAGE EPOCH: 66 LOSS:  2580.126220703125\n",
      "IMAGE EPOCH: 67 LOSS:  2579.80712890625\n",
      "IMAGE EPOCH: 68 LOSS:  2579.30810546875\n",
      "IMAGE EPOCH: 69 LOSS:  2578.61962890625\n",
      "IMAGE EPOCH: 70 LOSS:  2578.04345703125\n",
      "IMAGE EPOCH: 71 LOSS:  2577.70068359375\n",
      "IMAGE EPOCH: 72 LOSS:  2577.383544921875\n",
      "IMAGE EPOCH: 73 LOSS:  2576.90673828125\n",
      "IMAGE EPOCH: 74 LOSS:  2576.35986328125\n",
      "IMAGE EPOCH: 75 LOSS:  2575.921630859375\n",
      "IMAGE EPOCH: 76 LOSS:  2575.58349609375\n",
      "IMAGE EPOCH: 77 LOSS:  2575.2001953125\n",
      "IMAGE EPOCH: 78 LOSS:  2574.729736328125\n",
      "IMAGE EPOCH: 79 LOSS:  2574.277099609375\n",
      "IMAGE EPOCH: 80 LOSS:  2573.912841796875\n",
      "IMAGE EPOCH: 81 LOSS:  2573.57421875\n",
      "IMAGE EPOCH: 82 LOSS:  2573.178955078125\n",
      "IMAGE EPOCH: 83 LOSS:  2572.74365234375\n",
      "IMAGE EPOCH: 84 LOSS:  2572.34326171875\n",
      "IMAGE EPOCH: 85 LOSS:  2571.99169921875\n",
      "IMAGE EPOCH: 86 LOSS:  2571.6328125\n",
      "IMAGE EPOCH: 87 LOSS:  2571.238037109375\n",
      "IMAGE EPOCH: 88 LOSS:  2570.841552734375\n",
      "IMAGE EPOCH: 89 LOSS:  2570.475830078125\n",
      "IMAGE EPOCH: 90 LOSS:  2570.125732421875\n",
      "IMAGE EPOCH: 91 LOSS:  2569.75537109375\n",
      "IMAGE EPOCH: 92 LOSS:  2569.368408203125\n",
      "IMAGE EPOCH: 93 LOSS:  2568.993408203125\n",
      "IMAGE EPOCH: 94 LOSS:  2568.638671875\n",
      "IMAGE EPOCH: 95 LOSS:  2568.28564453125\n",
      "IMAGE EPOCH: 96 LOSS:  2567.918701171875\n",
      "IMAGE EPOCH: 97 LOSS:  2567.550048828125\n",
      "IMAGE EPOCH: 98 LOSS:  2567.194091796875\n",
      "IMAGE EPOCH: 99 LOSS:  2566.846923828125\n",
      "BATCH: 4\n",
      "::: PREPROCESSING: 240\n",
      "::: PREPROCESSING: 250\n",
      "::: PREPROCESSING: 260\n",
      "::: PREPROCESSING: 270\n",
      "::: PREPROCESSING: 280\n",
      "::: PREPROCESSING: 290\n",
      "IMAGE EPOCH: 0 LOSS:  2792.943115234375\n",
      "IMAGE EPOCH: 1 LOSS:  2782.242919921875\n",
      "IMAGE EPOCH: 2 LOSS:  2774.914794921875\n",
      "IMAGE EPOCH: 3 LOSS:  2776.60498046875\n",
      "IMAGE EPOCH: 4 LOSS:  2778.9599609375\n",
      "IMAGE EPOCH: 5 LOSS:  2774.109619140625\n",
      "IMAGE EPOCH: 6 LOSS:  2765.698974609375\n",
      "IMAGE EPOCH: 7 LOSS:  2761.762451171875\n",
      "IMAGE EPOCH: 8 LOSS:  2762.79052734375\n",
      "IMAGE EPOCH: 9 LOSS:  2762.525390625\n",
      "IMAGE EPOCH: 10 LOSS:  2757.985107421875\n",
      "IMAGE EPOCH: 11 LOSS:  2752.97314453125\n",
      "IMAGE EPOCH: 12 LOSS:  2751.53466796875\n",
      "IMAGE EPOCH: 13 LOSS:  2752.257080078125\n",
      "IMAGE EPOCH: 14 LOSS:  2751.26123046875\n",
      "IMAGE EPOCH: 15 LOSS:  2748.03564453125\n",
      "IMAGE EPOCH: 16 LOSS:  2745.458984375\n",
      "IMAGE EPOCH: 17 LOSS:  2745.094482421875\n",
      "IMAGE EPOCH: 18 LOSS:  2745.234375\n",
      "IMAGE EPOCH: 19 LOSS:  2743.8505859375\n",
      "IMAGE EPOCH: 20 LOSS:  2741.43408203125\n",
      "IMAGE EPOCH: 21 LOSS:  2739.815185546875\n",
      "IMAGE EPOCH: 22 LOSS:  2739.416015625\n",
      "IMAGE EPOCH: 23 LOSS:  2738.9521484375\n",
      "IMAGE EPOCH: 24 LOSS:  2737.52099609375\n",
      "IMAGE EPOCH: 25 LOSS:  2735.762451171875\n",
      "IMAGE EPOCH: 26 LOSS:  2734.684814453125\n",
      "IMAGE EPOCH: 27 LOSS:  2734.24853515625\n",
      "IMAGE EPOCH: 28 LOSS:  2733.626708984375\n",
      "IMAGE EPOCH: 29 LOSS:  2732.46240234375\n",
      "IMAGE EPOCH: 30 LOSS:  2731.274169921875\n",
      "IMAGE EPOCH: 31 LOSS:  2730.5771484375\n",
      "IMAGE EPOCH: 32 LOSS:  2730.193603515625\n",
      "IMAGE EPOCH: 33 LOSS:  2729.61328125\n",
      "IMAGE EPOCH: 34 LOSS:  2728.744873046875\n",
      "IMAGE EPOCH: 35 LOSS:  2727.942138671875\n",
      "IMAGE EPOCH: 36 LOSS:  2727.42822265625\n",
      "IMAGE EPOCH: 37 LOSS:  2727.01171875\n",
      "IMAGE EPOCH: 38 LOSS:  2726.41845703125\n",
      "IMAGE EPOCH: 39 LOSS:  2725.672607421875\n",
      "IMAGE EPOCH: 40 LOSS:  2725.0126953125\n",
      "IMAGE EPOCH: 41 LOSS:  2724.5283203125\n",
      "IMAGE EPOCH: 42 LOSS:  2724.064208984375\n",
      "IMAGE EPOCH: 43 LOSS:  2723.481689453125\n",
      "IMAGE EPOCH: 44 LOSS:  2722.83740234375\n",
      "IMAGE EPOCH: 45 LOSS:  2722.272705078125\n",
      "IMAGE EPOCH: 46 LOSS:  2721.802490234375\n",
      "IMAGE EPOCH: 47 LOSS:  2721.31591796875\n",
      "IMAGE EPOCH: 48 LOSS:  2720.75390625\n",
      "IMAGE EPOCH: 49 LOSS:  2720.176025390625\n",
      "IMAGE EPOCH: 50 LOSS:  2719.66064453125\n",
      "IMAGE EPOCH: 51 LOSS:  2719.194091796875\n",
      "IMAGE EPOCH: 52 LOSS:  2718.70556640625\n",
      "IMAGE EPOCH: 53 LOSS:  2718.174072265625\n",
      "IMAGE EPOCH: 54 LOSS:  2717.647216796875\n",
      "IMAGE EPOCH: 55 LOSS:  2717.16064453125\n",
      "IMAGE EPOCH: 56 LOSS:  2716.694580078125\n",
      "IMAGE EPOCH: 57 LOSS:  2716.207275390625\n",
      "IMAGE EPOCH: 58 LOSS:  2715.700439453125\n",
      "IMAGE EPOCH: 59 LOSS:  2715.205810546875\n",
      "IMAGE EPOCH: 60 LOSS:  2714.740966796875\n",
      "IMAGE EPOCH: 61 LOSS:  2714.283935546875\n",
      "IMAGE EPOCH: 62 LOSS:  2713.81591796875\n",
      "IMAGE EPOCH: 63 LOSS:  2713.345458984375\n",
      "IMAGE EPOCH: 64 LOSS:  2712.89208984375\n",
      "IMAGE EPOCH: 65 LOSS:  2712.456787109375\n",
      "IMAGE EPOCH: 66 LOSS:  2712.0224609375\n",
      "IMAGE EPOCH: 67 LOSS:  2711.580322265625\n",
      "IMAGE EPOCH: 68 LOSS:  2711.1416015625\n",
      "IMAGE EPOCH: 69 LOSS:  2710.714111328125\n",
      "IMAGE EPOCH: 70 LOSS:  2710.295654296875\n",
      "IMAGE EPOCH: 71 LOSS:  2709.875244140625\n",
      "IMAGE EPOCH: 72 LOSS:  2709.45166015625\n",
      "IMAGE EPOCH: 73 LOSS:  2709.032470703125\n",
      "IMAGE EPOCH: 74 LOSS:  2708.621337890625\n",
      "IMAGE EPOCH: 75 LOSS:  2708.21484375\n",
      "IMAGE EPOCH: 76 LOSS:  2707.806884765625\n",
      "IMAGE EPOCH: 77 LOSS:  2707.39892578125\n",
      "IMAGE EPOCH: 78 LOSS:  2706.996337890625\n",
      "IMAGE EPOCH: 79 LOSS:  2706.599853515625\n",
      "IMAGE EPOCH: 80 LOSS:  2706.20703125\n",
      "IMAGE EPOCH: 81 LOSS:  2705.81494140625\n",
      "IMAGE EPOCH: 82 LOSS:  2705.426025390625\n",
      "IMAGE EPOCH: 83 LOSS:  2705.04345703125\n",
      "IMAGE EPOCH: 84 LOSS:  2704.665771484375\n",
      "IMAGE EPOCH: 85 LOSS:  2704.2900390625\n",
      "IMAGE EPOCH: 86 LOSS:  2703.9169921875\n",
      "IMAGE EPOCH: 87 LOSS:  2703.548583984375\n",
      "IMAGE EPOCH: 88 LOSS:  2703.18359375\n",
      "IMAGE EPOCH: 89 LOSS:  2702.823486328125\n",
      "IMAGE EPOCH: 90 LOSS:  2702.462890625\n",
      "IMAGE EPOCH: 91 LOSS:  2702.105224609375\n",
      "IMAGE EPOCH: 92 LOSS:  2701.7509765625\n",
      "IMAGE EPOCH: 93 LOSS:  2701.3994140625\n",
      "IMAGE EPOCH: 94 LOSS:  2701.048828125\n",
      "IMAGE EPOCH: 95 LOSS:  2700.699462890625\n",
      "IMAGE EPOCH: 96 LOSS:  2700.352294921875\n",
      "IMAGE EPOCH: 97 LOSS:  2700.00537109375\n",
      "IMAGE EPOCH: 98 LOSS:  2699.659912109375\n",
      "IMAGE EPOCH: 99 LOSS:  2699.315185546875\n",
      "BATCH: 5\n",
      "::: PREPROCESSING: 300\n",
      "::: PREPROCESSING: 310\n",
      "::: PREPROCESSING: 320\n",
      "::: PREPROCESSING: 330\n",
      "::: PREPROCESSING: 340\n",
      "::: PREPROCESSING: 350\n",
      "IMAGE EPOCH: 0 LOSS:  3791.60400390625\n",
      "IMAGE EPOCH: 1 LOSS:  3630.9111328125\n",
      "IMAGE EPOCH: 2 LOSS:  3576.60107421875\n",
      "IMAGE EPOCH: 3 LOSS:  3406.09912109375\n",
      "IMAGE EPOCH: 4 LOSS:  3320.92724609375\n",
      "IMAGE EPOCH: 5 LOSS:  3349.36279296875\n",
      "IMAGE EPOCH: 6 LOSS:  3327.3046875\n",
      "IMAGE EPOCH: 7 LOSS:  3394.41845703125\n",
      "IMAGE EPOCH: 8 LOSS:  3416.875732421875\n",
      "IMAGE EPOCH: 9 LOSS:  3374.0732421875\n",
      "IMAGE EPOCH: 10 LOSS:  3364.201904296875\n",
      "IMAGE EPOCH: 11 LOSS:  3289.3515625\n",
      "IMAGE EPOCH: 12 LOSS:  3241.71533203125\n",
      "IMAGE EPOCH: 13 LOSS:  3232.49755859375\n",
      "IMAGE EPOCH: 14 LOSS:  3200.83203125\n",
      "IMAGE EPOCH: 15 LOSS:  3220.431884765625\n",
      "IMAGE EPOCH: 16 LOSS:  3230.52783203125\n",
      "IMAGE EPOCH: 17 LOSS:  3223.17041015625\n",
      "IMAGE EPOCH: 18 LOSS:  3231.810546875\n",
      "IMAGE EPOCH: 19 LOSS:  3206.4521484375\n",
      "IMAGE EPOCH: 20 LOSS:  3188.690673828125\n",
      "IMAGE EPOCH: 21 LOSS:  3178.2509765625\n",
      "IMAGE EPOCH: 22 LOSS:  3157.910888671875\n",
      "IMAGE EPOCH: 23 LOSS:  3163.022216796875\n",
      "IMAGE EPOCH: 24 LOSS:  3163.48828125\n",
      "IMAGE EPOCH: 25 LOSS:  3166.02734375\n",
      "IMAGE EPOCH: 26 LOSS:  3175.690185546875\n",
      "IMAGE EPOCH: 27 LOSS:  3169.4521484375\n",
      "IMAGE EPOCH: 28 LOSS:  3168.11474609375\n",
      "IMAGE EPOCH: 29 LOSS:  3161.064453125\n",
      "IMAGE EPOCH: 30 LOSS:  3150.2958984375\n",
      "IMAGE EPOCH: 31 LOSS:  3148.207275390625\n",
      "IMAGE EPOCH: 32 LOSS:  3142.03466796875\n",
      "IMAGE EPOCH: 33 LOSS:  3142.347900390625\n",
      "IMAGE EPOCH: 34 LOSS:  3144.182861328125\n",
      "IMAGE EPOCH: 35 LOSS:  3142.478271484375\n",
      "IMAGE EPOCH: 36 LOSS:  3144.254638671875\n",
      "IMAGE EPOCH: 37 LOSS:  3140.940185546875\n",
      "IMAGE EPOCH: 38 LOSS:  3137.6767578125\n",
      "IMAGE EPOCH: 39 LOSS:  3135.240234375\n",
      "IMAGE EPOCH: 40 LOSS:  3130.532470703125\n",
      "IMAGE EPOCH: 41 LOSS:  3129.525146484375\n",
      "IMAGE EPOCH: 42 LOSS:  3127.8720703125\n",
      "IMAGE EPOCH: 43 LOSS:  3126.966796875\n",
      "IMAGE EPOCH: 44 LOSS:  3127.339599609375\n",
      "IMAGE EPOCH: 45 LOSS:  3125.548583984375\n",
      "IMAGE EPOCH: 46 LOSS:  3124.6728515625\n",
      "IMAGE EPOCH: 47 LOSS:  3122.517822265625\n",
      "IMAGE EPOCH: 48 LOSS:  3119.90087890625\n",
      "IMAGE EPOCH: 49 LOSS:  3118.289306640625\n",
      "IMAGE EPOCH: 50 LOSS:  3115.94091796875\n",
      "IMAGE EPOCH: 51 LOSS:  3114.8349609375\n",
      "IMAGE EPOCH: 52 LOSS:  3113.75\n",
      "IMAGE EPOCH: 53 LOSS:  3112.51318359375\n",
      "IMAGE EPOCH: 54 LOSS:  3111.765625\n",
      "IMAGE EPOCH: 55 LOSS:  3110.1201171875\n",
      "IMAGE EPOCH: 56 LOSS:  3108.71044921875\n",
      "IMAGE EPOCH: 57 LOSS:  3107.0947265625\n",
      "IMAGE EPOCH: 58 LOSS:  3105.27978515625\n",
      "IMAGE EPOCH: 59 LOSS:  3103.993408203125\n",
      "IMAGE EPOCH: 60 LOSS:  3102.50927734375\n",
      "IMAGE EPOCH: 61 LOSS:  3101.45263671875\n",
      "IMAGE EPOCH: 62 LOSS:  3100.534912109375\n",
      "IMAGE EPOCH: 63 LOSS:  3099.553955078125\n",
      "IMAGE EPOCH: 64 LOSS:  3098.769287109375\n",
      "IMAGE EPOCH: 65 LOSS:  3097.64013671875\n",
      "IMAGE EPOCH: 66 LOSS:  3096.51904296875\n",
      "IMAGE EPOCH: 67 LOSS:  3095.28564453125\n",
      "IMAGE EPOCH: 68 LOSS:  3093.9677734375\n",
      "IMAGE EPOCH: 69 LOSS:  3092.86083984375\n",
      "IMAGE EPOCH: 70 LOSS:  3091.718994140625\n",
      "IMAGE EPOCH: 71 LOSS:  3090.754638671875\n",
      "IMAGE EPOCH: 72 LOSS:  3089.810302734375\n",
      "IMAGE EPOCH: 73 LOSS:  3088.798095703125\n",
      "IMAGE EPOCH: 74 LOSS:  3087.80908203125\n",
      "IMAGE EPOCH: 75 LOSS:  3086.68359375\n",
      "IMAGE EPOCH: 76 LOSS:  3085.583251953125\n",
      "IMAGE EPOCH: 77 LOSS:  3084.479248046875\n",
      "IMAGE EPOCH: 78 LOSS:  3083.40234375\n",
      "IMAGE EPOCH: 79 LOSS:  3082.438232421875\n",
      "IMAGE EPOCH: 80 LOSS:  3081.474609375\n",
      "IMAGE EPOCH: 81 LOSS:  3080.564208984375\n",
      "IMAGE EPOCH: 82 LOSS:  3079.6240234375\n",
      "IMAGE EPOCH: 83 LOSS:  3078.64599609375\n",
      "IMAGE EPOCH: 84 LOSS:  3077.67578125\n",
      "IMAGE EPOCH: 85 LOSS:  3076.680908203125\n",
      "IMAGE EPOCH: 86 LOSS:  3075.73583984375\n",
      "IMAGE EPOCH: 87 LOSS:  3074.809814453125\n",
      "IMAGE EPOCH: 88 LOSS:  3073.90380859375\n",
      "IMAGE EPOCH: 89 LOSS:  3073.018310546875\n",
      "IMAGE EPOCH: 90 LOSS:  3072.1171875\n",
      "IMAGE EPOCH: 91 LOSS:  3071.2333984375\n",
      "IMAGE EPOCH: 92 LOSS:  3070.34033203125\n",
      "IMAGE EPOCH: 93 LOSS:  3069.4541015625\n",
      "IMAGE EPOCH: 94 LOSS:  3068.582275390625\n",
      "IMAGE EPOCH: 95 LOSS:  3067.707763671875\n",
      "IMAGE EPOCH: 96 LOSS:  3066.850830078125\n",
      "IMAGE EPOCH: 97 LOSS:  3065.988037109375\n",
      "IMAGE EPOCH: 98 LOSS:  3065.12841796875\n",
      "IMAGE EPOCH: 99 LOSS:  3064.2705078125\n",
      "BATCH: 6\n",
      "::: PREPROCESSING: 360\n",
      "::: PREPROCESSING: 370\n",
      "::: PREPROCESSING: 380\n",
      "::: PREPROCESSING: 390\n",
      "::: PREPROCESSING: 400\n",
      "::: PREPROCESSING: 410\n",
      "IMAGE EPOCH: 0 LOSS:  4733.65478515625\n",
      "IMAGE EPOCH: 1 LOSS:  4656.34521484375\n",
      "IMAGE EPOCH: 2 LOSS:  4030.623046875\n",
      "IMAGE EPOCH: 3 LOSS:  3156.471923828125\n",
      "IMAGE EPOCH: 4 LOSS:  4316.28564453125\n",
      "IMAGE EPOCH: 5 LOSS:  2234.2802734375\n",
      "IMAGE EPOCH: 6 LOSS:  3986.52734375\n",
      "IMAGE EPOCH: 7 LOSS:  2493.862548828125\n",
      "IMAGE EPOCH: 8 LOSS:  2853.88232421875\n",
      "IMAGE EPOCH: 9 LOSS:  3148.330078125\n",
      "IMAGE EPOCH: 10 LOSS:  2090.23291015625\n",
      "IMAGE EPOCH: 11 LOSS:  3038.28662109375\n",
      "IMAGE EPOCH: 12 LOSS:  2391.623291015625\n",
      "IMAGE EPOCH: 13 LOSS:  2242.9921875\n",
      "IMAGE EPOCH: 14 LOSS:  2766.15234375\n",
      "IMAGE EPOCH: 15 LOSS:  2061.658447265625\n",
      "IMAGE EPOCH: 16 LOSS:  2401.566162109375\n",
      "IMAGE EPOCH: 17 LOSS:  2427.56103515625\n",
      "IMAGE EPOCH: 18 LOSS:  2019.230224609375\n",
      "IMAGE EPOCH: 19 LOSS:  2407.6923828125\n",
      "IMAGE EPOCH: 20 LOSS:  2200.637939453125\n",
      "IMAGE EPOCH: 21 LOSS:  2063.6953125\n",
      "IMAGE EPOCH: 22 LOSS:  2325.147216796875\n",
      "IMAGE EPOCH: 23 LOSS:  2085.16455078125\n",
      "IMAGE EPOCH: 24 LOSS:  2095.04638671875\n",
      "IMAGE EPOCH: 25 LOSS:  2230.24951171875\n",
      "IMAGE EPOCH: 26 LOSS:  2024.647216796875\n",
      "IMAGE EPOCH: 27 LOSS:  2102.005859375\n",
      "IMAGE EPOCH: 28 LOSS:  2145.124267578125\n",
      "IMAGE EPOCH: 29 LOSS:  1998.9814453125\n",
      "IMAGE EPOCH: 30 LOSS:  2084.7333984375\n",
      "IMAGE EPOCH: 31 LOSS:  2077.178466796875\n",
      "IMAGE EPOCH: 32 LOSS:  1985.81591796875\n",
      "IMAGE EPOCH: 33 LOSS:  2061.54638671875\n",
      "IMAGE EPOCH: 34 LOSS:  2026.951171875\n",
      "IMAGE EPOCH: 35 LOSS:  1978.07666015625\n",
      "IMAGE EPOCH: 36 LOSS:  2038.19580078125\n",
      "IMAGE EPOCH: 37 LOSS:  1993.65087890625\n",
      "IMAGE EPOCH: 38 LOSS:  1975.9290771484375\n",
      "IMAGE EPOCH: 39 LOSS:  2015.4896240234375\n",
      "IMAGE EPOCH: 40 LOSS:  1973.226806640625\n",
      "IMAGE EPOCH: 41 LOSS:  1975.1707763671875\n",
      "IMAGE EPOCH: 42 LOSS:  1995.85498046875\n",
      "IMAGE EPOCH: 43 LOSS:  1960.3668212890625\n",
      "IMAGE EPOCH: 44 LOSS:  1973.2630615234375\n",
      "IMAGE EPOCH: 45 LOSS:  1978.6759033203125\n",
      "IMAGE EPOCH: 46 LOSS:  1953.59375\n",
      "IMAGE EPOCH: 47 LOSS:  1969.072021484375\n",
      "IMAGE EPOCH: 48 LOSS:  1964.683349609375\n",
      "IMAGE EPOCH: 49 LOSS:  1950.363037109375\n",
      "IMAGE EPOCH: 50 LOSS:  1963.697021484375\n",
      "IMAGE EPOCH: 51 LOSS:  1954.1544189453125\n",
      "IMAGE EPOCH: 52 LOSS:  1948.4007568359375\n",
      "IMAGE EPOCH: 53 LOSS:  1957.4027099609375\n",
      "IMAGE EPOCH: 54 LOSS:  1946.70849609375\n",
      "IMAGE EPOCH: 55 LOSS:  1946.592529296875\n",
      "IMAGE EPOCH: 56 LOSS:  1950.541015625\n",
      "IMAGE EPOCH: 57 LOSS:  1941.569091796875\n",
      "IMAGE EPOCH: 58 LOSS:  1944.341064453125\n",
      "IMAGE EPOCH: 59 LOSS:  1944.2252197265625\n",
      "IMAGE EPOCH: 60 LOSS:  1938.1103515625\n",
      "IMAGE EPOCH: 61 LOSS:  1941.610107421875\n",
      "IMAGE EPOCH: 62 LOSS:  1939.08447265625\n",
      "IMAGE EPOCH: 63 LOSS:  1935.939697265625\n",
      "IMAGE EPOCH: 64 LOSS:  1938.5718994140625\n",
      "IMAGE EPOCH: 65 LOSS:  1935.1463623046875\n",
      "IMAGE EPOCH: 66 LOSS:  1934.337646484375\n",
      "IMAGE EPOCH: 67 LOSS:  1935.467529296875\n",
      "IMAGE EPOCH: 68 LOSS:  1932.27001953125\n",
      "IMAGE EPOCH: 69 LOSS:  1932.7054443359375\n",
      "IMAGE EPOCH: 70 LOSS:  1932.48388671875\n",
      "IMAGE EPOCH: 71 LOSS:  1930.17236328125\n",
      "IMAGE EPOCH: 72 LOSS:  1930.9281005859375\n",
      "IMAGE EPOCH: 73 LOSS:  1929.8115234375\n",
      "IMAGE EPOCH: 74 LOSS:  1928.5242919921875\n",
      "IMAGE EPOCH: 75 LOSS:  1929.027587890625\n",
      "IMAGE EPOCH: 76 LOSS:  1927.573486328125\n",
      "IMAGE EPOCH: 77 LOSS:  1927.0562744140625\n",
      "IMAGE EPOCH: 78 LOSS:  1927.06787109375\n",
      "IMAGE EPOCH: 79 LOSS:  1925.7412109375\n",
      "IMAGE EPOCH: 80 LOSS:  1925.631103515625\n",
      "IMAGE EPOCH: 81 LOSS:  1925.1771240234375\n",
      "IMAGE EPOCH: 82 LOSS:  1924.1976318359375\n",
      "IMAGE EPOCH: 83 LOSS:  1924.18017578125\n",
      "IMAGE EPOCH: 84 LOSS:  1923.4556884765625\n",
      "IMAGE EPOCH: 85 LOSS:  1922.839599609375\n",
      "IMAGE EPOCH: 86 LOSS:  1922.7012939453125\n",
      "IMAGE EPOCH: 87 LOSS:  1921.9287109375\n",
      "IMAGE EPOCH: 88 LOSS:  1921.5706787109375\n",
      "IMAGE EPOCH: 89 LOSS:  1921.2376708984375\n",
      "IMAGE EPOCH: 90 LOSS:  1920.5584716796875\n",
      "IMAGE EPOCH: 91 LOSS:  1920.3106689453125\n",
      "IMAGE EPOCH: 92 LOSS:  1919.83935546875\n",
      "IMAGE EPOCH: 93 LOSS:  1919.324951171875\n",
      "IMAGE EPOCH: 94 LOSS:  1919.0849609375\n",
      "IMAGE EPOCH: 95 LOSS:  1918.57568359375\n",
      "IMAGE EPOCH: 96 LOSS:  1918.1968994140625\n",
      "IMAGE EPOCH: 97 LOSS:  1917.8809814453125\n",
      "IMAGE EPOCH: 98 LOSS:  1917.394775390625\n",
      "IMAGE EPOCH: 99 LOSS:  1917.0941162109375\n",
      "BATCH: 7\n",
      "::: PREPROCESSING: 420\n",
      "::: PREPROCESSING: 430\n",
      "::: PREPROCESSING: 440\n",
      "::: PREPROCESSING: 450\n",
      "::: PREPROCESSING: 460\n",
      "::: PREPROCESSING: 470\n",
      "IMAGE EPOCH: 0 LOSS:  2565.0146484375\n",
      "IMAGE EPOCH: 1 LOSS:  2179.19091796875\n",
      "IMAGE EPOCH: 2 LOSS:  2525.19580078125\n",
      "IMAGE EPOCH: 3 LOSS:  2298.4765625\n",
      "IMAGE EPOCH: 4 LOSS:  2225.40576171875\n",
      "IMAGE EPOCH: 5 LOSS:  2410.41943359375\n",
      "IMAGE EPOCH: 6 LOSS:  2155.060546875\n",
      "IMAGE EPOCH: 7 LOSS:  2237.08154296875\n",
      "IMAGE EPOCH: 8 LOSS:  2311.76220703125\n",
      "IMAGE EPOCH: 9 LOSS:  2118.10302734375\n",
      "IMAGE EPOCH: 10 LOSS:  2222.00732421875\n",
      "IMAGE EPOCH: 11 LOSS:  2206.662353515625\n",
      "IMAGE EPOCH: 12 LOSS:  2103.08056640625\n",
      "IMAGE EPOCH: 13 LOSS:  2193.982421875\n",
      "IMAGE EPOCH: 14 LOSS:  2128.28466796875\n",
      "IMAGE EPOCH: 15 LOSS:  2085.66259765625\n",
      "IMAGE EPOCH: 16 LOSS:  2157.2763671875\n",
      "IMAGE EPOCH: 17 LOSS:  2087.761962890625\n",
      "IMAGE EPOCH: 18 LOSS:  2082.217041015625\n",
      "IMAGE EPOCH: 19 LOSS:  2122.698486328125\n",
      "IMAGE EPOCH: 20 LOSS:  2067.052001953125\n",
      "IMAGE EPOCH: 21 LOSS:  2085.070556640625\n",
      "IMAGE EPOCH: 22 LOSS:  2093.35546875\n",
      "IMAGE EPOCH: 23 LOSS:  2052.465087890625\n",
      "IMAGE EPOCH: 24 LOSS:  2078.810791015625\n",
      "IMAGE EPOCH: 25 LOSS:  2072.126953125\n",
      "IMAGE EPOCH: 26 LOSS:  2046.99365234375\n",
      "IMAGE EPOCH: 27 LOSS:  2069.163818359375\n",
      "IMAGE EPOCH: 28 LOSS:  2057.078125\n",
      "IMAGE EPOCH: 29 LOSS:  2047.6474609375\n",
      "IMAGE EPOCH: 30 LOSS:  2062.688232421875\n",
      "IMAGE EPOCH: 31 LOSS:  2047.4451904296875\n",
      "IMAGE EPOCH: 32 LOSS:  2046.885009765625\n",
      "IMAGE EPOCH: 33 LOSS:  2055.6650390625\n",
      "IMAGE EPOCH: 34 LOSS:  2041.657958984375\n",
      "IMAGE EPOCH: 35 LOSS:  2044.696044921875\n",
      "IMAGE EPOCH: 36 LOSS:  2048.072021484375\n",
      "IMAGE EPOCH: 37 LOSS:  2038.514404296875\n",
      "IMAGE EPOCH: 38 LOSS:  2043.3134765625\n",
      "IMAGE EPOCH: 39 LOSS:  2042.0814208984375\n",
      "IMAGE EPOCH: 40 LOSS:  2036.4398193359375\n",
      "IMAGE EPOCH: 41 LOSS:  2041.443115234375\n",
      "IMAGE EPOCH: 42 LOSS:  2037.89990234375\n",
      "IMAGE EPOCH: 43 LOSS:  2034.814453125\n",
      "IMAGE EPOCH: 44 LOSS:  2038.1741943359375\n",
      "IMAGE EPOCH: 45 LOSS:  2034.3922119140625\n",
      "IMAGE EPOCH: 46 LOSS:  2033.5985107421875\n",
      "IMAGE EPOCH: 47 LOSS:  2035.0294189453125\n",
      "IMAGE EPOCH: 48 LOSS:  2031.71826171875\n",
      "IMAGE EPOCH: 49 LOSS:  2032.467529296875\n",
      "IMAGE EPOCH: 50 LOSS:  2032.58203125\n",
      "IMAGE EPOCH: 51 LOSS:  2029.984130859375\n",
      "IMAGE EPOCH: 52 LOSS:  2031.0101318359375\n",
      "IMAGE EPOCH: 53 LOSS:  2030.31201171875\n",
      "IMAGE EPOCH: 54 LOSS:  2028.686279296875\n",
      "IMAGE EPOCH: 55 LOSS:  2029.40869140625\n",
      "IMAGE EPOCH: 56 LOSS:  2028.2054443359375\n",
      "IMAGE EPOCH: 57 LOSS:  2027.47509765625\n",
      "IMAGE EPOCH: 58 LOSS:  2027.884033203125\n",
      "IMAGE EPOCH: 59 LOSS:  2026.576171875\n",
      "IMAGE EPOCH: 60 LOSS:  2026.304443359375\n",
      "IMAGE EPOCH: 61 LOSS:  2026.3099365234375\n",
      "IMAGE EPOCH: 62 LOSS:  2025.2308349609375\n",
      "IMAGE EPOCH: 63 LOSS:  2025.16162109375\n",
      "IMAGE EPOCH: 64 LOSS:  2024.7779541015625\n",
      "IMAGE EPOCH: 65 LOSS:  2023.99072265625\n",
      "IMAGE EPOCH: 66 LOSS:  2023.9990234375\n",
      "IMAGE EPOCH: 67 LOSS:  2023.425048828125\n",
      "IMAGE EPOCH: 68 LOSS:  2022.8887939453125\n",
      "IMAGE EPOCH: 69 LOSS:  2022.842041015625\n",
      "IMAGE EPOCH: 70 LOSS:  2022.2657470703125\n",
      "IMAGE EPOCH: 71 LOSS:  2021.925537109375\n",
      "IMAGE EPOCH: 72 LOSS:  2021.7327880859375\n",
      "IMAGE EPOCH: 73 LOSS:  2021.205322265625\n",
      "IMAGE EPOCH: 74 LOSS:  2020.9898681640625\n",
      "IMAGE EPOCH: 75 LOSS:  2020.6871337890625\n",
      "IMAGE EPOCH: 76 LOSS:  2020.22705078125\n",
      "IMAGE EPOCH: 77 LOSS:  2020.040283203125\n",
      "IMAGE EPOCH: 78 LOSS:  2019.693359375\n",
      "IMAGE EPOCH: 79 LOSS:  2019.3304443359375\n",
      "IMAGE EPOCH: 80 LOSS:  2019.119384765625\n",
      "IMAGE EPOCH: 81 LOSS:  2018.7562255859375\n",
      "IMAGE EPOCH: 82 LOSS:  2018.4737548828125\n",
      "IMAGE EPOCH: 83 LOSS:  2018.237060546875\n",
      "IMAGE EPOCH: 84 LOSS:  2017.8948974609375\n",
      "IMAGE EPOCH: 85 LOSS:  2017.665283203125\n",
      "IMAGE EPOCH: 86 LOSS:  2017.4169921875\n",
      "IMAGE EPOCH: 87 LOSS:  2017.1236572265625\n",
      "IMAGE EPOCH: 88 LOSS:  2016.9176025390625\n",
      "IMAGE EPOCH: 89 LOSS:  2016.6644287109375\n",
      "IMAGE EPOCH: 90 LOSS:  2016.419921875\n",
      "IMAGE EPOCH: 91 LOSS:  2016.225830078125\n",
      "IMAGE EPOCH: 92 LOSS:  2015.9837646484375\n",
      "IMAGE EPOCH: 93 LOSS:  2015.77783203125\n",
      "IMAGE EPOCH: 94 LOSS:  2015.5880126953125\n",
      "IMAGE EPOCH: 95 LOSS:  2015.368408203125\n",
      "IMAGE EPOCH: 96 LOSS:  2015.186279296875\n",
      "IMAGE EPOCH: 97 LOSS:  2014.9959716796875\n",
      "IMAGE EPOCH: 98 LOSS:  2014.79931640625\n",
      "IMAGE EPOCH: 99 LOSS:  2014.6309814453125\n",
      "BATCH: 8\n",
      "::: PREPROCESSING: 480\n",
      "::: PREPROCESSING: 490\n",
      "::: PREPROCESSING: 500\n",
      "::: PREPROCESSING: 510\n",
      "::: PREPROCESSING: 520\n",
      "::: PREPROCESSING: 530\n",
      "IMAGE EPOCH: 0 LOSS:  6671.55908203125\n",
      "IMAGE EPOCH: 1 LOSS:  6328.25341796875\n",
      "IMAGE EPOCH: 2 LOSS:  6114.3671875\n",
      "IMAGE EPOCH: 3 LOSS:  6223.11328125\n",
      "IMAGE EPOCH: 4 LOSS:  6345.54931640625\n",
      "IMAGE EPOCH: 5 LOSS:  6220.9541015625\n",
      "IMAGE EPOCH: 6 LOSS:  6015.85986328125\n",
      "IMAGE EPOCH: 7 LOSS:  5977.2470703125\n",
      "IMAGE EPOCH: 8 LOSS:  6067.314453125\n",
      "IMAGE EPOCH: 9 LOSS:  6093.83251953125\n",
      "IMAGE EPOCH: 10 LOSS:  5999.9853515625\n",
      "IMAGE EPOCH: 11 LOSS:  5905.53466796875\n",
      "IMAGE EPOCH: 12 LOSS:  5916.69921875\n",
      "IMAGE EPOCH: 13 LOSS:  5985.8818359375\n",
      "IMAGE EPOCH: 14 LOSS:  5995.814453125\n",
      "IMAGE EPOCH: 15 LOSS:  5927.8544921875\n",
      "IMAGE EPOCH: 16 LOSS:  5868.095703125\n",
      "IMAGE EPOCH: 17 LOSS:  5872.13232421875\n",
      "IMAGE EPOCH: 18 LOSS:  5901.96240234375\n",
      "IMAGE EPOCH: 19 LOSS:  5896.8798828125\n",
      "IMAGE EPOCH: 20 LOSS:  5855.13330078125\n",
      "IMAGE EPOCH: 21 LOSS:  5823.92626953125\n",
      "IMAGE EPOCH: 22 LOSS:  5829.03662109375\n",
      "IMAGE EPOCH: 23 LOSS:  5843.6455078125\n",
      "IMAGE EPOCH: 24 LOSS:  5832.96240234375\n",
      "IMAGE EPOCH: 25 LOSS:  5803.35888671875\n",
      "IMAGE EPOCH: 26 LOSS:  5786.6494140625\n",
      "IMAGE EPOCH: 27 LOSS:  5792.55322265625\n",
      "IMAGE EPOCH: 28 LOSS:  5800.056640625\n",
      "IMAGE EPOCH: 29 LOSS:  5790.8115234375\n",
      "IMAGE EPOCH: 30 LOSS:  5772.2578125\n",
      "IMAGE EPOCH: 31 LOSS:  5763.2685546875\n",
      "IMAGE EPOCH: 32 LOSS:  5767.5732421875\n",
      "IMAGE EPOCH: 33 LOSS:  5771.6591796875\n",
      "IMAGE EPOCH: 34 LOSS:  5765.80224609375\n",
      "IMAGE EPOCH: 35 LOSS:  5756.02001953125\n",
      "IMAGE EPOCH: 36 LOSS:  5752.552734375\n",
      "IMAGE EPOCH: 37 LOSS:  5755.1376953125\n",
      "IMAGE EPOCH: 38 LOSS:  5755.4326171875\n",
      "IMAGE EPOCH: 39 LOSS:  5749.68115234375\n",
      "IMAGE EPOCH: 40 LOSS:  5742.8828125\n",
      "IMAGE EPOCH: 41 LOSS:  5740.62548828125\n",
      "IMAGE EPOCH: 42 LOSS:  5741.50244140625\n",
      "IMAGE EPOCH: 43 LOSS:  5740.1689453125\n",
      "IMAGE EPOCH: 44 LOSS:  5735.17626953125\n",
      "IMAGE EPOCH: 45 LOSS:  5730.2509765625\n",
      "IMAGE EPOCH: 46 LOSS:  5728.4619140625\n",
      "IMAGE EPOCH: 47 LOSS:  5728.35009765625\n",
      "IMAGE EPOCH: 48 LOSS:  5726.7685546875\n",
      "IMAGE EPOCH: 49 LOSS:  5723.3134765625\n",
      "IMAGE EPOCH: 50 LOSS:  5720.23681640625\n",
      "IMAGE EPOCH: 51 LOSS:  5718.89453125\n",
      "IMAGE EPOCH: 52 LOSS:  5718.10546875\n",
      "IMAGE EPOCH: 53 LOSS:  5716.20166015625\n",
      "IMAGE EPOCH: 54 LOSS:  5713.37060546875\n",
      "IMAGE EPOCH: 55 LOSS:  5711.072265625\n",
      "IMAGE EPOCH: 56 LOSS:  5709.8505859375\n",
      "IMAGE EPOCH: 57 LOSS:  5708.767578125\n",
      "IMAGE EPOCH: 58 LOSS:  5706.9501953125\n",
      "IMAGE EPOCH: 59 LOSS:  5704.7578125\n",
      "IMAGE EPOCH: 60 LOSS:  5703.08203125\n",
      "IMAGE EPOCH: 61 LOSS:  5702.0625\n",
      "IMAGE EPOCH: 62 LOSS:  5701.0146484375\n",
      "IMAGE EPOCH: 63 LOSS:  5699.46484375\n",
      "IMAGE EPOCH: 64 LOSS:  5697.7294921875\n",
      "IMAGE EPOCH: 65 LOSS:  5696.32080078125\n",
      "IMAGE EPOCH: 66 LOSS:  5695.21923828125\n",
      "IMAGE EPOCH: 67 LOSS:  5694.005859375\n",
      "IMAGE EPOCH: 68 LOSS:  5692.51708984375\n",
      "IMAGE EPOCH: 69 LOSS:  5691.02490234375\n",
      "IMAGE EPOCH: 70 LOSS:  5689.81494140625\n",
      "IMAGE EPOCH: 71 LOSS:  5688.77294921875\n",
      "IMAGE EPOCH: 72 LOSS:  5687.61474609375\n",
      "IMAGE EPOCH: 73 LOSS:  5686.2890625\n",
      "IMAGE EPOCH: 74 LOSS:  5684.986328125\n",
      "IMAGE EPOCH: 75 LOSS:  5683.82470703125\n",
      "IMAGE EPOCH: 76 LOSS:  5682.6884765625\n",
      "IMAGE EPOCH: 77 LOSS:  5681.42724609375\n",
      "IMAGE EPOCH: 78 LOSS:  5680.0673828125\n",
      "IMAGE EPOCH: 79 LOSS:  5678.75341796875\n",
      "IMAGE EPOCH: 80 LOSS:  5677.52978515625\n",
      "IMAGE EPOCH: 81 LOSS:  5676.31396484375\n",
      "IMAGE EPOCH: 82 LOSS:  5675.03857421875\n",
      "IMAGE EPOCH: 83 LOSS:  5673.75146484375\n",
      "IMAGE EPOCH: 84 LOSS:  5672.52978515625\n",
      "IMAGE EPOCH: 85 LOSS:  5671.3681640625\n",
      "IMAGE EPOCH: 86 LOSS:  5670.2109375\n",
      "IMAGE EPOCH: 87 LOSS:  5669.03857421875\n",
      "IMAGE EPOCH: 88 LOSS:  5667.89208984375\n",
      "IMAGE EPOCH: 89 LOSS:  5666.80615234375\n",
      "IMAGE EPOCH: 90 LOSS:  5665.755859375\n",
      "IMAGE EPOCH: 91 LOSS:  5664.69970703125\n",
      "IMAGE EPOCH: 92 LOSS:  5663.63232421875\n",
      "IMAGE EPOCH: 93 LOSS:  5662.580078125\n",
      "IMAGE EPOCH: 94 LOSS:  5661.55078125\n",
      "IMAGE EPOCH: 95 LOSS:  5660.521484375\n",
      "IMAGE EPOCH: 96 LOSS:  5659.4697265625\n",
      "IMAGE EPOCH: 97 LOSS:  5658.4013671875\n",
      "IMAGE EPOCH: 98 LOSS:  5657.33837890625\n",
      "IMAGE EPOCH: 99 LOSS:  5656.2802734375\n",
      "BATCH: 9\n",
      "::: PREPROCESSING: 540\n",
      "::: PREPROCESSING: 550\n",
      "::: PREPROCESSING: 560\n",
      "::: PREPROCESSING: 570\n",
      "::: PREPROCESSING: 580\n",
      "::: PREPROCESSING: 590\n",
      "IMAGE EPOCH: 0 LOSS:  4485.20849609375\n",
      "IMAGE EPOCH: 1 LOSS:  4350.1767578125\n",
      "IMAGE EPOCH: 2 LOSS:  4292.416015625\n",
      "IMAGE EPOCH: 3 LOSS:  4235.3330078125\n",
      "IMAGE EPOCH: 4 LOSS:  4102.8115234375\n",
      "IMAGE EPOCH: 5 LOSS:  4024.49365234375\n",
      "IMAGE EPOCH: 6 LOSS:  4011.997802734375\n",
      "IMAGE EPOCH: 7 LOSS:  3952.419921875\n",
      "IMAGE EPOCH: 8 LOSS:  3895.344970703125\n",
      "IMAGE EPOCH: 9 LOSS:  3909.9599609375\n",
      "IMAGE EPOCH: 10 LOSS:  3916.812255859375\n",
      "IMAGE EPOCH: 11 LOSS:  3893.66064453125\n",
      "IMAGE EPOCH: 12 LOSS:  3907.270751953125\n",
      "IMAGE EPOCH: 13 LOSS:  3928.9091796875\n",
      "IMAGE EPOCH: 14 LOSS:  3912.5966796875\n",
      "IMAGE EPOCH: 15 LOSS:  3904.62744140625\n",
      "IMAGE EPOCH: 16 LOSS:  3913.74365234375\n",
      "IMAGE EPOCH: 17 LOSS:  3899.001708984375\n",
      "IMAGE EPOCH: 18 LOSS:  3879.334228515625\n",
      "IMAGE EPOCH: 19 LOSS:  3879.6484375\n",
      "IMAGE EPOCH: 20 LOSS:  3872.9609375\n",
      "IMAGE EPOCH: 21 LOSS:  3855.62646484375\n",
      "IMAGE EPOCH: 22 LOSS:  3852.086669921875\n",
      "IMAGE EPOCH: 23 LOSS:  3852.56982421875\n",
      "IMAGE EPOCH: 24 LOSS:  3842.4296875\n",
      "IMAGE EPOCH: 25 LOSS:  3836.400390625\n",
      "IMAGE EPOCH: 26 LOSS:  3838.260009765625\n",
      "IMAGE EPOCH: 27 LOSS:  3833.87109375\n",
      "IMAGE EPOCH: 28 LOSS:  3827.332275390625\n",
      "IMAGE EPOCH: 29 LOSS:  3828.065185546875\n",
      "IMAGE EPOCH: 30 LOSS:  3827.931640625\n",
      "IMAGE EPOCH: 31 LOSS:  3823.399169921875\n",
      "IMAGE EPOCH: 32 LOSS:  3822.69970703125\n",
      "IMAGE EPOCH: 33 LOSS:  3823.974853515625\n",
      "IMAGE EPOCH: 34 LOSS:  3821.11328125\n",
      "IMAGE EPOCH: 35 LOSS:  3818.57373046875\n",
      "IMAGE EPOCH: 36 LOSS:  3818.54833984375\n",
      "IMAGE EPOCH: 37 LOSS:  3816.083740234375\n",
      "IMAGE EPOCH: 38 LOSS:  3812.211669921875\n",
      "IMAGE EPOCH: 39 LOSS:  3810.746337890625\n",
      "IMAGE EPOCH: 40 LOSS:  3809.073486328125\n",
      "IMAGE EPOCH: 41 LOSS:  3805.944580078125\n",
      "IMAGE EPOCH: 42 LOSS:  3804.520263671875\n",
      "IMAGE EPOCH: 43 LOSS:  3804.124755859375\n",
      "IMAGE EPOCH: 44 LOSS:  3802.529052734375\n",
      "IMAGE EPOCH: 45 LOSS:  3801.389892578125\n",
      "IMAGE EPOCH: 46 LOSS:  3801.328125\n",
      "IMAGE EPOCH: 47 LOSS:  3800.38916015625\n",
      "IMAGE EPOCH: 48 LOSS:  3799.013671875\n",
      "IMAGE EPOCH: 49 LOSS:  3798.39111328125\n",
      "IMAGE EPOCH: 50 LOSS:  3797.3935546875\n",
      "IMAGE EPOCH: 51 LOSS:  3795.763671875\n",
      "IMAGE EPOCH: 52 LOSS:  3794.66552734375\n",
      "IMAGE EPOCH: 53 LOSS:  3793.71923828125\n",
      "IMAGE EPOCH: 54 LOSS:  3792.3076171875\n",
      "IMAGE EPOCH: 55 LOSS:  3791.194091796875\n",
      "IMAGE EPOCH: 56 LOSS:  3790.447265625\n",
      "IMAGE EPOCH: 57 LOSS:  3789.420654296875\n",
      "IMAGE EPOCH: 58 LOSS:  3788.43701171875\n",
      "IMAGE EPOCH: 59 LOSS:  3787.798095703125\n",
      "IMAGE EPOCH: 60 LOSS:  3787.0205078125\n",
      "IMAGE EPOCH: 61 LOSS:  3786.12548828125\n",
      "IMAGE EPOCH: 62 LOSS:  3785.46826171875\n",
      "IMAGE EPOCH: 63 LOSS:  3784.7822265625\n",
      "IMAGE EPOCH: 64 LOSS:  3783.933349609375\n",
      "IMAGE EPOCH: 65 LOSS:  3783.216796875\n",
      "IMAGE EPOCH: 66 LOSS:  3782.560791015625\n",
      "IMAGE EPOCH: 67 LOSS:  3781.781005859375\n",
      "IMAGE EPOCH: 68 LOSS:  3781.056884765625\n",
      "IMAGE EPOCH: 69 LOSS:  3780.4365234375\n",
      "IMAGE EPOCH: 70 LOSS:  3779.74755859375\n",
      "IMAGE EPOCH: 71 LOSS:  3779.064697265625\n",
      "IMAGE EPOCH: 72 LOSS:  3778.4755859375\n",
      "IMAGE EPOCH: 73 LOSS:  3777.85888671875\n",
      "IMAGE EPOCH: 74 LOSS:  3777.21484375\n",
      "IMAGE EPOCH: 75 LOSS:  3776.63427734375\n",
      "IMAGE EPOCH: 76 LOSS:  3776.05322265625\n",
      "IMAGE EPOCH: 77 LOSS:  3775.430419921875\n",
      "IMAGE EPOCH: 78 LOSS:  3774.84033203125\n",
      "IMAGE EPOCH: 79 LOSS:  3774.262451171875\n",
      "IMAGE EPOCH: 80 LOSS:  3773.64892578125\n",
      "IMAGE EPOCH: 81 LOSS:  3773.049072265625\n",
      "IMAGE EPOCH: 82 LOSS:  3772.472412109375\n",
      "IMAGE EPOCH: 83 LOSS:  3771.8759765625\n",
      "IMAGE EPOCH: 84 LOSS:  3771.2841796875\n",
      "IMAGE EPOCH: 85 LOSS:  3770.717041015625\n",
      "IMAGE EPOCH: 86 LOSS:  3770.140869140625\n",
      "IMAGE EPOCH: 87 LOSS:  3769.559814453125\n",
      "IMAGE EPOCH: 88 LOSS:  3768.995361328125\n",
      "IMAGE EPOCH: 89 LOSS:  3768.4267578125\n",
      "IMAGE EPOCH: 90 LOSS:  3767.85009765625\n",
      "IMAGE EPOCH: 91 LOSS:  3767.284912109375\n",
      "IMAGE EPOCH: 92 LOSS:  3766.720703125\n",
      "IMAGE EPOCH: 93 LOSS:  3766.151611328125\n",
      "IMAGE EPOCH: 94 LOSS:  3765.5908203125\n",
      "IMAGE EPOCH: 95 LOSS:  3765.03466796875\n",
      "IMAGE EPOCH: 96 LOSS:  3764.47412109375\n",
      "IMAGE EPOCH: 97 LOSS:  3763.91796875\n",
      "IMAGE EPOCH: 98 LOSS:  3763.365966796875\n",
      "IMAGE EPOCH: 99 LOSS:  3762.811279296875\n",
      "BATCH: 10\n",
      "::: PREPROCESSING: 600\n",
      "::: PREPROCESSING: 610\n",
      "::: PREPROCESSING: 620\n",
      "::: PREPROCESSING: 630\n",
      "::: PREPROCESSING: 640\n",
      "::: PREPROCESSING: 650\n",
      "IMAGE EPOCH: 0 LOSS:  5371.4208984375\n",
      "IMAGE EPOCH: 1 LOSS:  4189.7431640625\n",
      "IMAGE EPOCH: 2 LOSS:  4388.90869140625\n",
      "IMAGE EPOCH: 3 LOSS:  4981.205078125\n",
      "IMAGE EPOCH: 4 LOSS:  4500.12109375\n",
      "IMAGE EPOCH: 5 LOSS:  4152.5576171875\n",
      "IMAGE EPOCH: 6 LOSS:  4431.06787109375\n",
      "IMAGE EPOCH: 7 LOSS:  4448.912109375\n",
      "IMAGE EPOCH: 8 LOSS:  4167.8037109375\n",
      "IMAGE EPOCH: 9 LOSS:  4150.76904296875\n",
      "IMAGE EPOCH: 10 LOSS:  4259.171875\n",
      "IMAGE EPOCH: 11 LOSS:  4135.97998046875\n",
      "IMAGE EPOCH: 12 LOSS:  4007.09814453125\n",
      "IMAGE EPOCH: 13 LOSS:  4098.32568359375\n",
      "IMAGE EPOCH: 14 LOSS:  4135.37255859375\n",
      "IMAGE EPOCH: 15 LOSS:  3980.96533203125\n",
      "IMAGE EPOCH: 16 LOSS:  3925.177001953125\n",
      "IMAGE EPOCH: 17 LOSS:  4026.9697265625\n",
      "IMAGE EPOCH: 18 LOSS:  4023.286865234375\n",
      "IMAGE EPOCH: 19 LOSS:  3893.7822265625\n",
      "IMAGE EPOCH: 20 LOSS:  3870.500732421875\n",
      "IMAGE EPOCH: 21 LOSS:  3958.6025390625\n",
      "IMAGE EPOCH: 22 LOSS:  3948.3935546875\n",
      "IMAGE EPOCH: 23 LOSS:  3845.611572265625\n",
      "IMAGE EPOCH: 24 LOSS:  3829.820556640625\n",
      "IMAGE EPOCH: 25 LOSS:  3889.99267578125\n",
      "IMAGE EPOCH: 26 LOSS:  3880.968017578125\n",
      "IMAGE EPOCH: 27 LOSS:  3814.1455078125\n",
      "IMAGE EPOCH: 28 LOSS:  3803.630859375\n",
      "IMAGE EPOCH: 29 LOSS:  3836.05078125\n",
      "IMAGE EPOCH: 30 LOSS:  3824.121337890625\n",
      "IMAGE EPOCH: 31 LOSS:  3785.52734375\n",
      "IMAGE EPOCH: 32 LOSS:  3781.91796875\n",
      "IMAGE EPOCH: 33 LOSS:  3792.57373046875\n",
      "IMAGE EPOCH: 34 LOSS:  3778.0009765625\n",
      "IMAGE EPOCH: 35 LOSS:  3760.40087890625\n",
      "IMAGE EPOCH: 36 LOSS:  3761.491455078125\n",
      "IMAGE EPOCH: 37 LOSS:  3758.17919921875\n",
      "IMAGE EPOCH: 38 LOSS:  3743.56005859375\n",
      "IMAGE EPOCH: 39 LOSS:  3739.1279296875\n",
      "IMAGE EPOCH: 40 LOSS:  3740.1767578125\n",
      "IMAGE EPOCH: 41 LOSS:  3729.019775390625\n",
      "IMAGE EPOCH: 42 LOSS:  3717.96826171875\n",
      "IMAGE EPOCH: 43 LOSS:  3720.14111328125\n",
      "IMAGE EPOCH: 44 LOSS:  3718.490234375\n",
      "IMAGE EPOCH: 45 LOSS:  3705.158203125\n",
      "IMAGE EPOCH: 46 LOSS:  3699.010009765625\n",
      "IMAGE EPOCH: 47 LOSS:  3702.22900390625\n",
      "IMAGE EPOCH: 48 LOSS:  3697.203125\n",
      "IMAGE EPOCH: 49 LOSS:  3686.364013671875\n",
      "IMAGE EPOCH: 50 LOSS:  3684.0830078125\n",
      "IMAGE EPOCH: 51 LOSS:  3684.33984375\n",
      "IMAGE EPOCH: 52 LOSS:  3677.25439453125\n",
      "IMAGE EPOCH: 53 LOSS:  3670.783935546875\n",
      "IMAGE EPOCH: 54 LOSS:  3669.575439453125\n",
      "IMAGE EPOCH: 55 LOSS:  3666.12646484375\n",
      "IMAGE EPOCH: 56 LOSS:  3660.0927734375\n",
      "IMAGE EPOCH: 57 LOSS:  3656.880859375\n",
      "IMAGE EPOCH: 58 LOSS:  3654.07275390625\n",
      "IMAGE EPOCH: 59 LOSS:  3649.209716796875\n",
      "IMAGE EPOCH: 60 LOSS:  3645.73583984375\n",
      "IMAGE EPOCH: 61 LOSS:  3643.18408203125\n",
      "IMAGE EPOCH: 62 LOSS:  3638.818359375\n",
      "IMAGE EPOCH: 63 LOSS:  3635.087890625\n",
      "IMAGE EPOCH: 64 LOSS:  3633.102294921875\n",
      "IMAGE EPOCH: 65 LOSS:  3629.685546875\n",
      "IMAGE EPOCH: 66 LOSS:  3625.56884765625\n",
      "IMAGE EPOCH: 67 LOSS:  3623.311767578125\n",
      "IMAGE EPOCH: 68 LOSS:  3620.890380859375\n",
      "IMAGE EPOCH: 69 LOSS:  3617.1064453125\n",
      "IMAGE EPOCH: 70 LOSS:  3614.234130859375\n",
      "IMAGE EPOCH: 71 LOSS:  3612.049072265625\n",
      "IMAGE EPOCH: 72 LOSS:  3608.938720703125\n",
      "IMAGE EPOCH: 73 LOSS:  3605.9619140625\n",
      "IMAGE EPOCH: 74 LOSS:  3603.630615234375\n",
      "IMAGE EPOCH: 75 LOSS:  3600.912109375\n",
      "IMAGE EPOCH: 76 LOSS:  3598.1787109375\n",
      "IMAGE EPOCH: 77 LOSS:  3595.84326171875\n",
      "IMAGE EPOCH: 78 LOSS:  3593.271484375\n",
      "IMAGE EPOCH: 79 LOSS:  3590.74951171875\n",
      "IMAGE EPOCH: 80 LOSS:  3588.5732421875\n",
      "IMAGE EPOCH: 81 LOSS:  3586.119873046875\n",
      "IMAGE EPOCH: 82 LOSS:  3583.619140625\n",
      "IMAGE EPOCH: 83 LOSS:  3581.500244140625\n",
      "IMAGE EPOCH: 84 LOSS:  3579.227294921875\n",
      "IMAGE EPOCH: 85 LOSS:  3576.819091796875\n",
      "IMAGE EPOCH: 86 LOSS:  3574.748291015625\n",
      "IMAGE EPOCH: 87 LOSS:  3572.663330078125\n",
      "IMAGE EPOCH: 88 LOSS:  3570.43603515625\n",
      "IMAGE EPOCH: 89 LOSS:  3568.408203125\n",
      "IMAGE EPOCH: 90 LOSS:  3566.4033203125\n",
      "IMAGE EPOCH: 91 LOSS:  3564.319580078125\n",
      "IMAGE EPOCH: 92 LOSS:  3562.353515625\n",
      "IMAGE EPOCH: 93 LOSS:  3560.38818359375\n",
      "IMAGE EPOCH: 94 LOSS:  3558.39990234375\n",
      "IMAGE EPOCH: 95 LOSS:  3556.517578125\n",
      "IMAGE EPOCH: 96 LOSS:  3554.609375\n",
      "IMAGE EPOCH: 97 LOSS:  3552.698486328125\n",
      "IMAGE EPOCH: 98 LOSS:  3550.89697265625\n",
      "IMAGE EPOCH: 99 LOSS:  3549.07177734375\n"
     ]
    }
   ],
   "source": [
    "N_query_points = 150\n",
    "EPOCHS = 3\n",
    "IMAGE_EPOCHS = 100\n",
    "\n",
    "for epoch in range(EPOCHS):\n",
    "    match epoch:\n",
    "        case 0:\n",
    "            batches = get_batched_viewpoint(total=len(nsr.images),\n",
    "                                            step=20,\n",
    "                                            batch_size=24, \n",
    "                                            epochs=1)[0]\n",
    "        case 1:\n",
    "            batches = get_batched_viewpoint(total=len(nsr.images),\n",
    "                                            step=10,\n",
    "                                            batch_size=12, \n",
    "                                            epochs=1)[0]\n",
    "            \n",
    "            nsr.double_grid()\n",
    "            nsr.increment_sharpness()\n",
    "            N_query_points *= 2\n",
    "        case 2:\n",
    "            batches = get_batched_viewpoint(total=len(nsr.images),\n",
    "                                            step=10,\n",
    "                                            batch_size=6, \n",
    "                                            epochs=1)[0]\n",
    "            \n",
    "            nsr.double_grid()\n",
    "            nsr.increment_sharpness()\n",
    "            N_query_points *= 2\n",
    "\n",
    "    print(\"EPOCH:\", epoch, \"SHARPNESS:\", nsr.model.sharpness, \"FG VOXEL SIZE:\", nsr.fg_V.voxel_size)\n",
    "\n",
    "    for batch, viewpoints in enumerate(batches):\n",
    "        print(\"BATCH:\", batch)\n",
    "\n",
    "        btp, bst, bv = [], [], []\n",
    "        for viewpoint in viewpoints:\n",
    "            print(\"::: PREPROCESSING:\", viewpoint)\n",
    "            \n",
    "            tp, st, v = nsr.query_data(viewpoint, N_query_points=N_query_points)\n",
    "            \n",
    "            btp.append(tp[torch.newaxis])\n",
    "            bst.append(st[torch.newaxis])\n",
    "            bv.append(v[torch.newaxis])\n",
    "        \n",
    "        btp = torch.concat(tuple(btp))\n",
    "        bst = torch.concat(tuple(bst))\n",
    "        bv = torch.concat(tuple(bv))\n",
    "\n",
    "        for image_epoch in range(0, IMAGE_EPOCHS):\n",
    "            print(\"IMAGE EPOCH:\", image_epoch, end=\"\\t\")\n",
    "\n",
    "            optimizer.zero_grad()\n",
    "\n",
    "            loss, _ = nsr.model(btp, bst, bv)\n",
    "            loss.backward()\n",
    "            optimizer.step()\n",
    "\n",
    "            # nsr.update_sharpness()\n",
    "        \n",
    "            print(\"LOSS: \", loss.item())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x17601b86ec0>"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tp, st, v = nsr.query_data(200, N_query_points=600)\n",
    "_, rendering = nsr.model(tp[torch.newaxis], st[torch.newaxis], v[torch.newaxis])\n",
    "\n",
    "plt.imshow(tp.detach().cpu().reshape((1080//16+1, 1920//16, 3)).permute(1,0,2).float().numpy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x17601f68400>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "img = rendering.cpu().reshape((1080//16+1, 1920//16, 3)).permute(1,0,2)\n",
    "\n",
    "plt.imshow(img.detach().numpy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save(nsr.model.state_dict(), \"./nsr000.pt\")\n",
    "torch.save(optimizer.state_dict(), \"./nsr000o.pt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([1, 1, 32, 32, 32]), torch.Size([1, 1, 32, 32, 32]))"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nsr.model.s.fg_sdf.shape, nsr.model.s.bg_sdf.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "::: VIEWPOINT: 0\n",
      "BATCH: 0\n",
      "BATCH: 1\n",
      "BATCH: 2\n",
      "BATCH: 3\n",
      "BATCH: 4\n",
      "BATCH: 5\n",
      "BATCH: 6\n",
      "BATCH: 7\n",
      "BATCH: 8\n",
      "BATCH: 9\n",
      "BATCH: 10\n",
      "BATCH: 11\n",
      "BATCH: 12\n",
      "BATCH: 13\n",
      "BATCH: 14\n",
      "BATCH: 15\n",
      "BATCH: 16\n",
      "BATCH: 17\n",
      "BATCH: 18\n",
      "BATCH: 19\n",
      "BATCH: 20\n",
      "BATCH: 21\n",
      "BATCH: 22\n",
      "BATCH: 23\n",
      "IMAGE EPOCH: 0\n",
      "139263.515625\n",
      "IMAGE EPOCH: 1\n",
      "104403.8515625\n",
      "IMAGE EPOCH: 2\n",
      "82899.8828125\n",
      "IMAGE EPOCH: 3\n",
      "72281.3515625\n",
      "IMAGE EPOCH: 4\n",
      "70233.8125\n",
      "IMAGE EPOCH: 5\n",
      "70079.015625\n",
      "IMAGE EPOCH: 6\n",
      "66544.8203125\n",
      "IMAGE EPOCH: 7\n",
      "59744.27734375\n",
      "IMAGE EPOCH: 8\n",
      "51911.3359375\n",
      "IMAGE EPOCH: 9\n",
      "45096.265625\n",
      "IMAGE EPOCH: 10\n",
      "40257.33984375\n",
      "IMAGE EPOCH: 11\n",
      "37128.078125\n",
      "IMAGE EPOCH: 12\n",
      "34865.76171875\n",
      "IMAGE EPOCH: 13\n",
      "32705.115234375\n",
      "IMAGE EPOCH: 14\n",
      "30330.291015625\n",
      "IMAGE EPOCH: 15\n",
      "27862.94140625\n",
      "IMAGE EPOCH: 16\n",
      "25691.46875\n",
      "IMAGE EPOCH: 17\n",
      "24167.697265625\n",
      "IMAGE EPOCH: 18\n",
      "23250.466796875\n",
      "IMAGE EPOCH: 19\n",
      "22432.27734375\n",
      "IMAGE EPOCH: 20\n",
      "21211.880859375\n",
      "IMAGE EPOCH: 21\n",
      "19648.47265625\n",
      "IMAGE EPOCH: 22\n",
      "18220.962890625\n",
      "IMAGE EPOCH: 23\n",
      "17288.98046875\n",
      "IMAGE EPOCH: 24\n",
      "16790.03515625\n",
      "IMAGE EPOCH: 25\n",
      "16349.2900390625\n",
      "IMAGE EPOCH: 26\n",
      "15629.296875\n",
      "IMAGE EPOCH: 27\n",
      "14611.572265625\n",
      "IMAGE EPOCH: 28\n",
      "13548.0263671875\n",
      "IMAGE EPOCH: 29\n",
      "12779.505859375\n",
      "IMAGE EPOCH: 30\n",
      "12380.841796875\n",
      "IMAGE EPOCH: 31\n",
      "12066.51953125\n",
      "IMAGE EPOCH: 32\n",
      "11567.306640625\n",
      "IMAGE EPOCH: 33\n",
      "10937.90625\n",
      "IMAGE EPOCH: 34\n",
      "10404.927734375\n",
      "IMAGE EPOCH: 35\n",
      "10041.927734375\n",
      "IMAGE EPOCH: 36\n",
      "9723.4443359375\n",
      "IMAGE EPOCH: 37\n",
      "9374.34375\n",
      "IMAGE EPOCH: 38\n",
      "9082.583984375\n",
      "IMAGE EPOCH: 39\n",
      "8907.0537109375\n",
      "IMAGE EPOCH: 40\n",
      "8730.4306640625\n",
      "IMAGE EPOCH: 41\n",
      "8473.6806640625\n",
      "IMAGE EPOCH: 42\n",
      "8244.7236328125\n",
      "IMAGE EPOCH: 43\n",
      "8122.99560546875\n",
      "IMAGE EPOCH: 44\n",
      "8022.06396484375\n",
      "IMAGE EPOCH: 45\n",
      "7864.5341796875\n",
      "IMAGE EPOCH: 46\n",
      "7690.16943359375\n",
      "IMAGE EPOCH: 47\n",
      "7534.3876953125\n",
      "IMAGE EPOCH: 48\n",
      "7391.52978515625\n",
      "IMAGE EPOCH: 49\n",
      "7281.2138671875\n",
      "IMAGE EPOCH: 50\n",
      "7203.5888671875\n",
      "IMAGE EPOCH: 51\n",
      "7107.513671875\n",
      "IMAGE EPOCH: 52\n",
      "6987.38330078125\n",
      "IMAGE EPOCH: 53\n",
      "6904.9580078125\n",
      "IMAGE EPOCH: 54\n",
      "6838.435546875\n",
      "IMAGE EPOCH: 55\n",
      "6786.66259765625\n",
      "IMAGE EPOCH: 56\n",
      "6742.333984375\n",
      "IMAGE EPOCH: 57\n",
      "6681.23486328125\n",
      "IMAGE EPOCH: 58\n",
      "6622.0634765625\n",
      "IMAGE EPOCH: 59\n",
      "6587.263671875\n",
      "IMAGE EPOCH: 60\n",
      "6555.30615234375\n",
      "IMAGE EPOCH: 61\n",
      "6507.3828125\n",
      "IMAGE EPOCH: 62\n",
      "6454.63427734375\n",
      "IMAGE EPOCH: 63\n",
      "6409.9521484375\n",
      "IMAGE EPOCH: 64\n",
      "6374.55859375\n",
      "IMAGE EPOCH: 65\n",
      "6341.9638671875\n",
      "IMAGE EPOCH: 66\n",
      "6297.41162109375\n",
      "IMAGE EPOCH: 67\n",
      "6250.5087890625\n",
      "IMAGE EPOCH: 68\n",
      "6220.064453125\n",
      "IMAGE EPOCH: 69\n",
      "6192.00927734375\n",
      "IMAGE EPOCH: 70\n",
      "6153.576171875\n",
      "IMAGE EPOCH: 71\n",
      "6116.326171875\n",
      "IMAGE EPOCH: 72\n",
      "6086.2490234375\n",
      "IMAGE EPOCH: 73\n",
      "6058.7880859375\n",
      "IMAGE EPOCH: 74\n",
      "6030.4345703125\n",
      "IMAGE EPOCH: 75\n",
      "6002.8935546875\n",
      "IMAGE EPOCH: 76\n",
      "5975.0791015625\n",
      "IMAGE EPOCH: 77\n",
      "5948.73193359375\n",
      "IMAGE EPOCH: 78\n",
      "5927.166015625\n",
      "IMAGE EPOCH: 79\n",
      "5905.40283203125\n",
      "IMAGE EPOCH: 80\n",
      "5881.66064453125\n",
      "IMAGE EPOCH: 81\n",
      "5859.7919921875\n",
      "IMAGE EPOCH: 82\n",
      "5839.318359375\n",
      "IMAGE EPOCH: 83\n",
      "5818.97412109375\n",
      "IMAGE EPOCH: 84\n",
      "5799.5361328125\n",
      "IMAGE EPOCH: 85\n",
      "5781.05322265625\n",
      "IMAGE EPOCH: 86\n",
      "5763.921875\n",
      "IMAGE EPOCH: 87\n",
      "5747.0\n",
      "IMAGE EPOCH: 88\n",
      "5729.79736328125\n",
      "IMAGE EPOCH: 89\n",
      "5712.65771484375\n",
      "IMAGE EPOCH: 90\n",
      "5696.4794921875\n",
      "IMAGE EPOCH: 91\n",
      "5680.55224609375\n",
      "IMAGE EPOCH: 92\n",
      "5664.521484375\n",
      "IMAGE EPOCH: 93\n",
      "5647.9736328125\n",
      "IMAGE EPOCH: 94\n",
      "5632.1689453125\n",
      "IMAGE EPOCH: 95\n",
      "5616.79052734375\n",
      "IMAGE EPOCH: 96\n",
      "5600.9228515625\n",
      "IMAGE EPOCH: 97\n",
      "5585.1240234375\n",
      "IMAGE EPOCH: 98\n",
      "5569.63037109375\n",
      "IMAGE EPOCH: 99\n",
      "5554.89453125\n"
     ]
    }
   ],
   "source": [
    "# torch.autograd.set_detect_anomaly(True)\n",
    "# torch.cuda.empty_cache()\n",
    "\n",
    "EPOCHS = 1\n",
    "IMAGE_EPOCHS = 100\n",
    "BATCH_SIZE = 24\n",
    "\n",
    "N_query_points = 150\n",
    "\n",
    "for epoch in range(0, EPOCHS):\n",
    "\n",
    "    for viewpoint in range(0, 1, BATCH_SIZE): #len(nsr.images)//10):\n",
    "        \n",
    "        print(f\"::: VIEWPOINT: {viewpoint}\")\n",
    "\n",
    "        # for g in optimizer.param_groups:\n",
    "        #             g[\"lr\"] = 0.01\n",
    "\n",
    "        # if 35 > viewpoint >= 30:\n",
    "        #     nsr.grid_doubling()\n",
    "        #     N_query_points *= 2\n",
    "\n",
    "        # HxWx3=Nx3, NxTx3, 2xN\n",
    "        btp, bst, bv = [], [], []\n",
    "        for b in range(BATCH_SIZE):\n",
    "            tp, st, v = nsr.query_data(viewpoint+b, N_query_points=N_query_points,\n",
    "                                    t_size=abs(nsr.fg_V.step_size/2))\n",
    "            \n",
    "            # print(tp.shape)\n",
    "            # print(st.shape)\n",
    "            # print(v.shape)\n",
    "            print(\"BATCH:\", viewpoint+b)\n",
    "            btp.append(tp[torch.newaxis])\n",
    "            bst.append(st[torch.newaxis])\n",
    "            bv.append(v[torch.newaxis])\n",
    "        \n",
    "        # btp = tp[torch.newaxis]\n",
    "        # bst = st[torch.newaxis]\n",
    "        # bv = v[torch.newaxis]\n",
    "        # btp = torch.tensor(btp)\n",
    "        # bst = torch.tensor(bst)\n",
    "        # bv = torch.tensor(bv)\n",
    "        btp = torch.concat(tuple(btp))\n",
    "        bst = torch.concat(tuple(bst))\n",
    "        bv = torch.concat(tuple(bv))\n",
    "\n",
    "        for image_epoch in range(0, IMAGE_EPOCHS):\n",
    "            print(f\"IMAGE EPOCH: {image_epoch}\")\n",
    "\n",
    "            total_loss = 0.0\n",
    "\n",
    "           \n",
    "            # if image_epoch == 50:\n",
    "            #     for g in optimizer.param_groups:\n",
    "            #         g[\"lr\"] = 0.001\n",
    "\n",
    "\n",
    "            # MINIBATCH_SIZE = st.shape[0]//4\n",
    "            # for i in range(0, st.shape[0], MINIBATCH_SIZE):\n",
    "            optimizer.zero_grad()\n",
    "\n",
    "            # output = nsr.model(st, v)\n",
    "            # output = output.cpu() #.detach()\n",
    "\n",
    "            # # print(output.requires_grad)\n",
    "\n",
    "            # # photo_loss = torch.linalg.norm(tp[i:i+MINIBATCH_SIZE]-output, dim=1).sum()\n",
    "            # # photo_loss.backward()\n",
    "\n",
    "            # loss = photo_loss(tp, output)\n",
    "            # loss.backward()\n",
    "\n",
    "            loss, _ = nsr.model(btp, bst, bv)\n",
    "            loss.backward()\n",
    "\n",
    "            total_loss += loss.item()\n",
    "\n",
    "            optimizer.step()\n",
    "        \n",
    "        \n",
    "            # scheduler.step(total_loss)\n",
    "\n",
    "            # nsr.update_sharpness()\n",
    "\n",
    "        \n",
    "            print(total_loss)\n",
    "            # print(scheduler.get_last_lr())\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.01\n"
     ]
    }
   ],
   "source": [
    "for g in optimizer.param_groups:\n",
    "    print(g[\"lr\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'st' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "Cell \u001b[1;32mIn[4], line 5\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorchview\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m draw_graph\n\u001b[0;32m      3\u001b[0m \u001b[38;5;66;03m# torchviz.make_dot(output, nsr.model.named_parameters())\u001b[39;00m\n\u001b[1;32m----> 5\u001b[0m model_graph \u001b[38;5;241m=\u001b[39m draw_graph(nsr\u001b[38;5;241m.\u001b[39mmodel, input_size\u001b[38;5;241m=\u001b[39m[\u001b[43mst\u001b[49m\u001b[38;5;241m.\u001b[39mshape, v\u001b[38;5;241m.\u001b[39mshape], depth\u001b[38;5;241m=\u001b[39m\u001b[38;5;241m1\u001b[39m, device\u001b[38;5;241m=\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mmeta\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n\u001b[0;32m      6\u001b[0m model_graph\u001b[38;5;241m.\u001b[39mvisual_graph\n",
      "\u001b[1;31mNameError\u001b[0m: name 'st' is not defined"
     ]
    }
   ],
   "source": [
    "from torchview import draw_graph\n",
    "\n",
    "# torchviz.make_dot(output, nsr.model.named_parameters())\n",
    "\n",
    "model_graph = draw_graph(nsr.model, input_size=[st.shape, v.shape], depth=1, device=\"meta\")\n",
    "model_graph.visual_graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'graphviz.graphs.Digraph'>\n"
     ]
    }
   ],
   "source": [
    "print(type(model_graph.visual_graph))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "(process:24424): Pango-WARNING **: 09:38:03.421: couldn't load font \"Linux libertine Not-Rotated 10\", falling back to \"Sans Not-Rotated 10\", expect ugly output.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'nsr_graph.png.pdf'"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_graph.visual_graph.render(\"./nsr_graph.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "(process:27400): Pango-WARNING **: 09:48:44.052: couldn't load font \"Linux libertine Not-Rotated 10\", falling back to \"Sans Not-Rotated 10\", expect ugly output.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'S_graph.png.pdf'"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_graph = draw_graph(nsr.model.s, input_size=(st.shape[0], st.shape[2]), depth=2, device=\"meta\")\n",
    "model_graph.visual_graph.render(\"./S_graph.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "(process:24908): Pango-WARNING **: 09:48:44.862: couldn't load font \"Linux libertine Not-Rotated 10\", falling back to \"Sans Not-Rotated 10\", expect ugly output.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'Lo_graph.png.pdf'"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_graph = draw_graph(nsr.model.lo, input_size=[(st.shape[0], st.shape[2]), v.shape], depth=2, device=\"meta\")\n",
    "model_graph.visual_graph.render(\"./Lo_graph.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# _, st, _ = nsr.query_data(50,\n",
    "#                           N_query_points=150, t_size=abs(nsr.fg_V.step_size/2),  # 0.15, #\n",
    "#                           _pixel_grid_width_step=25.0,\n",
    "#                           _pixel_grid_height_step=25.0,\n",
    "#                           _scale = 1)\n",
    "_, st, _ = nsr.query_data(0, N_query_points=300, t_size=abs(nsr.fg_V.step_size/2),\n",
    "                          _pixel_grid_width_step=10.0,\n",
    "                          _pixel_grid_height_step=10.0,)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([336, 300, 3]), tensor(-0.0484, device='cuda:0'))"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "st.shape, nsr.fg_V.step_size/2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(tensor([-1.7500,  1.5000,  0.5000], device='cuda:0'),\n",
       " tensor([1.5000, 1.5000, 1.5000], device='cuda:0'),\n",
       " tensor(-0.0968, device='cuda:0'),\n",
       " torch.Size([1, 1, 32, 32, 32]),\n",
       " torch.Size([1, 1, 32, 32, 32]))"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nsr.model.s.fg.center, nsr.model.s.fg.scale, nsr.fg_V.step_size, nsr.model.s.fg_sdf.shape, nsr.model.s.bg_sdf.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "go.Figure(data=[\n",
    "    go.Scatter3d(\n",
    "        x=nsr.fg_V.grid_pos[0].flatten().cpu().detach().numpy(),\n",
    "        y=nsr.fg_V.grid_pos[1].flatten().cpu().detach().numpy(),\n",
    "        z=nsr.fg_V.grid_pos[2].flatten().cpu().detach().numpy(),\n",
    "        mode='markers',\n",
    "        marker={\n",
    "            'size': 10,\n",
    "            \"color\": nsr.fg_V.sdf.flatten().cpu().detach().numpy(),\n",
    "            'opacity': 0.3,\n",
    "            \"cmax\": 1.0,\n",
    "            \"cmin\": -1.0,\n",
    "            \"colorbar\": {\n",
    "                \"thickness\": 10,\n",
    "            }\n",
    "        }\n",
    "    ),\n",
    "    # go.Scatter3d(\n",
    "    #     x=nsr.bg_V.grid_pos[0].flatten().cpu().detach().numpy(),\n",
    "    #     y=nsr.bg_V.grid_pos[1].flatten().cpu().detach().numpy(),\n",
    "    #     z=nsr.bg_V.grid_pos[2].flatten().cpu().detach().numpy(),\n",
    "    #     mode='markers',\n",
    "    #     marker={\n",
    "    #         'size': 10,\n",
    "    #         \"color\": nsr.bg_V.sdf.flatten().cpu().detach().numpy(),\n",
    "    #         'opacity': 0.2,\n",
    "    #         \"cmax\": 1.0,\n",
    "    #         \"cmin\": -1.0,\n",
    "    #         # \"colorbar\": {\n",
    "    #         #     \"thickness\": 10,\n",
    "    #         # }\n",
    "    #     }\n",
    "    # ),\n",
    "    # go.Scatter3d(\n",
    "    #     x=new_fg.grid_pos[0].flatten().cpu().detach().numpy(),\n",
    "    #     y=new_fg.grid_pos[1].flatten().cpu().detach().numpy(),\n",
    "    #     z=new_fg.grid_pos[2].flatten().cpu().detach().numpy(),\n",
    "    #     mode='markers',\n",
    "    #     marker={\n",
    "    #         'size': 10,\n",
    "    #         \"color\": new_fg.sdf.flatten().cpu().detach().numpy(),\n",
    "    #         'opacity': 0.3,\n",
    "    #         \"cmax\": 1.0,\n",
    "    #         \"cmin\": -1.0,\n",
    "    #         \"colorbar\": {\n",
    "    #             \"thickness\": 10,\n",
    "    #         }\n",
    "    #     }\n",
    "    # ),\n",
    "    # go.Scatter3d(\n",
    "    #     x=new_bg.grid_pos[0].flatten().cpu().detach().numpy(),\n",
    "    #     y=new_bg.grid_pos[1].flatten().cpu().detach().numpy(),\n",
    "    #     z=new_bg.grid_pos[2].flatten().cpu().detach().numpy(),\n",
    "    #     mode='markers',\n",
    "    #     marker={\n",
    "    #         'size': 10,\n",
    "    #         \"color\": new_bg.sdf.flatten().cpu().detach().numpy(),\n",
    "    #         'opacity': 0.2,\n",
    "    #         \"cmax\": 1.0,\n",
    "    #         \"cmin\": -1.0,\n",
    "    #         # \"colorbar\": {\n",
    "    #         #     \"thickness\": 10,\n",
    "    #         # }\n",
    "    #     }\n",
    "    # ),\n",
    "    # go.Scatter3d(\n",
    "    #     x=st[:,:,0].flatten().cpu().numpy(),\n",
    "    #     y=st[:,:,1].flatten().cpu().numpy(),\n",
    "    #     z=st[:,:,2].flatten().cpu().numpy(),\n",
    "    #     mode='markers',\n",
    "    #     marker={\n",
    "    #         'size': 3,\n",
    "    #         'opacity': 1.0,\n",
    "    #     }\n",
    "    # ),\n",
    "], layout=go.Layout(\n",
    "    margin={'l': 0, 'r': 0, 'b': 0, 't': 0},\n",
    "    scene=dict(\n",
    "        aspectmode=\"cube\",\n",
    "        xaxis=dict(range=[-30, 30]),  # Set x-axis range\n",
    "        yaxis=dict(range=[-30, 30]),  # Set y-axis range\n",
    "        zaxis=dict(range=[-30, 30])   # Set z-axis range\n",
    "    ),\n",
    "    scene_camera= dict(\n",
    "        eye=dict(x=0.0, y=0.0, z=0.1)\n",
    "    )\n",
    "))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}