train_check.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "3dc48f3f-5f59-4cc6-92d8-e6e4aed3c7aa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/teamspace/studios/this_studio/Deeplab\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/zeus/miniconda3/envs/cloudspace/lib/python3.10/site-packages/IPython/core/magics/osm.py:417: UserWarning: using dhist requires you to install the `pickleshare` library.\n",
      "  self.shell.db['dhist'] = compress_dhist(dhist)[-100:]\n"
     ]
    }
   ],
   "source": [
    "%cd Deeplab\n",
    "# Deeplab/saved_models/100_percent_volume_inhouse_resized_0.5432411918859545.pth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "316f2e67-ef4d-4af5-aab3-8486d27a0365",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/teamspace/studios/this_studio/Deeplab\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "print(os.getcwd())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "a1e0b12b-64d9-4da8-b6c4-4b135d6a4871",
   "metadata": {},
   "outputs": [],
   "source": [
    "!python3 split_dataset.py --dataset_dir 'crf_sam_annotations_large'\n",
    "# /teamspace/studios/this_studio/Deeplab/india_sam_dino_annotations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "597c0faf-0f21-4060-b660-f9946b6a4f87",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "!python3 split_dataset.py --dataset_dir 'pilot_annotations_eastafrica_haiti_freetown'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "4e55d553-85ff-48ca-942c-7c610a4c6550",
   "metadata": {},
   "outputs": [],
   "source": [
    "downstreamed_large = '/teamspace/studios/this_studio/Deeplab/saved_models/india_sam_dino_250_samples_0.8495187035894669.pth'\n",
    "downstreamed_small = '/teamspace/studios/this_studio/Deeplab/saved_models/india_sam_dino_250_samples_0.8940880817269634.pth'\n",
    "downstreamed_tuned = '/teamspace/studios/this_studio/Deeplab/saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8213798905479425.pth'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "6eda478a-c228-4ded-8ce0-308a744da310",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device: cuda\n",
      "Dataset: custom, Train set: 1860, Val set: 496\n",
      "[!] Retrain\n",
      "Epoch 1, Itrs 10/10000, Loss=1.270395\n",
      "Epoch 1, Itrs 20/10000, Loss=1.171225\n",
      "Epoch 1, Itrs 30/10000, Loss=1.158787\n",
      "Epoch 1, Itrs 40/10000, Loss=1.156444\n",
      "Epoch 1, Itrs 50/10000, Loss=1.087996\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.56it/s]\n",
      "Validation Loss: 0.705588\n",
      "\n",
      "Foreground Acc: 0.864146\n",
      "Precision: 0.764606\n",
      "Recall: 0.864146\n",
      "F1 Score: 0.811334\n",
      "IoU Foreground: 0.682559\n",
      "IoU Background: 0.896930\n",
      "Mean IoU: 0.789744\n",
      "\n",
      "[0.6825586295035073]\n",
      "[50]\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_without_prob0.6825586295035073.pth\n",
      "saved here.\n",
      "==========================================================\n",
      "Epoch 1, Itrs 60/10000, Loss=1.084551\n",
      "Epoch 1, Itrs 70/10000, Loss=1.049040\n",
      "Epoch 1, Itrs 80/10000, Loss=1.056198\n",
      "Epoch 1, Itrs 90/10000, Loss=0.978865\n",
      "Epoch 1, Itrs 100/10000, Loss=1.012903\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.46it/s]\n",
      "Validation Loss: 0.467086\n",
      "\n",
      "Foreground Acc: 0.940243\n",
      "Precision: 0.812285\n",
      "Recall: 0.940243\n",
      "F1 Score: 0.871593\n",
      "IoU Foreground: 0.772410\n",
      "IoU Background: 0.927535\n",
      "Mean IoU: 0.849973\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423]\n",
      "[50, 100]\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_without_prob0.7724098165632423.pth\n",
      "saved here.\n",
      "==========================================================\n",
      "Epoch 1, Itrs 110/10000, Loss=0.954925\n",
      "Epoch 1, Itrs 120/10000, Loss=1.000235\n",
      "Epoch 1, Itrs 130/10000, Loss=0.952086\n",
      "Epoch 1, Itrs 140/10000, Loss=1.019137\n",
      "Epoch 1, Itrs 150/10000, Loss=0.950899\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.44it/s]\n",
      "Validation Loss: 0.642673\n",
      "\n",
      "Foreground Acc: 0.849777\n",
      "Precision: 0.904482\n",
      "Recall: 0.849777\n",
      "F1 Score: 0.876277\n",
      "IoU Foreground: 0.779798\n",
      "IoU Background: 0.938685\n",
      "Mean IoU: 0.859241\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844]\n",
      "[50, 100, 150]\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_without_prob0.7797975776744844.pth\n",
      "saved here.\n",
      "==========================================================\n",
      "Epoch 1, Itrs 160/10000, Loss=1.001252\n",
      "Epoch 1, Itrs 170/10000, Loss=1.105563\n",
      "Epoch 1, Itrs 180/10000, Loss=0.991002\n",
      "Epoch 1, Itrs 190/10000, Loss=0.976842\n",
      "Epoch 1, Itrs 200/10000, Loss=0.973417\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.21it/s]\n",
      "Validation Loss: 0.667816\n",
      "\n",
      "Foreground Acc: 0.870258\n",
      "Precision: 0.884281\n",
      "Recall: 0.870258\n",
      "F1 Score: 0.877213\n",
      "IoU Foreground: 0.781282\n",
      "IoU Background: 0.937421\n",
      "Mean IoU: 0.859352\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834]\n",
      "[50, 100, 150, 200]\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_without_prob0.7812819754598834.pth\n",
      "saved here.\n",
      "==========================================================\n",
      "Epoch 1, Itrs 210/10000, Loss=0.956479\n",
      "Epoch 1, Itrs 220/10000, Loss=1.007021\n",
      "Epoch 1, Itrs 230/10000, Loss=0.957513\n",
      "Epoch 1, Itrs 240/10000, Loss=0.965469\n",
      "Epoch 1, Itrs 250/10000, Loss=0.933239\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.34it/s]\n",
      "Validation Loss: 0.591800\n",
      "\n",
      "Foreground Acc: 0.830356\n",
      "Precision: 0.915127\n",
      "Recall: 0.830356\n",
      "F1 Score: 0.870683\n",
      "IoU Foreground: 0.770982\n",
      "IoU Background: 0.937286\n",
      "Mean IoU: 0.854134\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267]\n",
      "[50, 100, 150, 200, 250]\n",
      "==========================================================\n",
      "Epoch 1, Itrs 260/10000, Loss=0.943957\n",
      "Epoch 1, Itrs 270/10000, Loss=0.912549\n",
      "Epoch 1, Itrs 280/10000, Loss=1.032744\n",
      "Epoch 1, Itrs 290/10000, Loss=0.909566\n",
      "Epoch 1, Itrs 300/10000, Loss=0.975351\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 15.90it/s]\n",
      "Validation Loss: 0.428412\n",
      "\n",
      "Foreground Acc: 0.960664\n",
      "Precision: 0.875346\n",
      "Recall: 0.960664\n",
      "F1 Score: 0.916023\n",
      "IoU Foreground: 0.845057\n",
      "IoU Background: 0.953680\n",
      "Mean IoU: 0.899369\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855]\n",
      "[50, 100, 150, 200, 250, 300]\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_without_prob0.8450568952991855.pth\n",
      "saved here.\n",
      "==========================================================\n",
      "Epoch 1, Itrs 310/10000, Loss=0.985129\n",
      "Epoch 1, Itrs 320/10000, Loss=1.035901\n",
      "Epoch 1, Itrs 330/10000, Loss=0.995695\n",
      "Epoch 1, Itrs 340/10000, Loss=0.976834\n",
      "Epoch 1, Itrs 350/10000, Loss=0.889982\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.47it/s]\n",
      "Validation Loss: 0.723050\n",
      "\n",
      "Foreground Acc: 0.984706\n",
      "Precision: 0.681064\n",
      "Recall: 0.984706\n",
      "F1 Score: 0.805211\n",
      "IoU Foreground: 0.673935\n",
      "IoU Background: 0.873918\n",
      "Mean IoU: 0.773926\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348]\n",
      "[50, 100, 150, 200, 250, 300, 350]\n",
      "==========================================================\n",
      "Epoch 1, Itrs 360/10000, Loss=0.883644\n",
      "Epoch 1, Itrs 370/10000, Loss=0.957049\n",
      "Epoch 1, Itrs 380/10000, Loss=0.925711\n",
      "Epoch 1, Itrs 390/10000, Loss=0.964440\n",
      "Epoch 1, Itrs 400/10000, Loss=0.956309\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.69it/s]\n",
      "Validation Loss: 0.504383\n",
      "\n",
      "Foreground Acc: 0.947903\n",
      "Precision: 0.855156\n",
      "Recall: 0.947903\n",
      "F1 Score: 0.899144\n",
      "IoU Foreground: 0.816768\n",
      "IoU Background: 0.944266\n",
      "Mean IoU: 0.880517\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400]\n",
      "==========================================================\n",
      "Epoch 1, Itrs 410/10000, Loss=0.919564\n",
      "Epoch 1, Itrs 420/10000, Loss=0.998073\n",
      "Epoch 1, Itrs 430/10000, Loss=0.930845\n",
      "Epoch 1, Itrs 440/10000, Loss=0.923193\n",
      "Epoch 1, Itrs 450/10000, Loss=0.929978\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.33it/s]\n",
      "Validation Loss: 0.555984\n",
      "\n",
      "Foreground Acc: 0.974538\n",
      "Precision: 0.822019\n",
      "Recall: 0.974538\n",
      "F1 Score: 0.891804\n",
      "IoU Foreground: 0.804735\n",
      "IoU Background: 0.937588\n",
      "Mean IoU: 0.871162\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450]\n",
      "==========================================================\n",
      "Epoch 1, Itrs 460/10000, Loss=0.999309\n",
      "Epoch 2, Itrs 470/10000, Loss=0.948554\n",
      "Epoch 2, Itrs 480/10000, Loss=0.969003\n",
      "Epoch 2, Itrs 490/10000, Loss=0.947140\n",
      "Epoch 2, Itrs 500/10000, Loss=0.962995\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.52it/s]\n",
      "Validation Loss: 0.515012\n",
      "\n",
      "Foreground Acc: 0.947138\n",
      "Precision: 0.892842\n",
      "Recall: 0.947138\n",
      "F1 Score: 0.919189\n",
      "IoU Foreground: 0.850462\n",
      "IoU Background: 0.956361\n",
      "Mean IoU: 0.903412\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500]\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_without_prob0.8504620350701382.pth\n",
      "saved here.\n",
      "==========================================================\n",
      "Epoch 2, Itrs 510/10000, Loss=0.885336\n",
      "Epoch 2, Itrs 520/10000, Loss=0.920848\n",
      "Epoch 2, Itrs 530/10000, Loss=0.887701\n",
      "Epoch 2, Itrs 540/10000, Loss=0.855330\n",
      "Epoch 2, Itrs 550/10000, Loss=0.918397\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.96it/s]\n",
      "Validation Loss: 0.469004\n",
      "\n",
      "Foreground Acc: 0.952572\n",
      "Precision: 0.909401\n",
      "Recall: 0.952572\n",
      "F1 Score: 0.930486\n",
      "IoU Foreground: 0.870008\n",
      "IoU Background: 0.962651\n",
      "Mean IoU: 0.916330\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550]\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_without_prob0.8700079479483994.pth\n",
      "saved here.\n",
      "==========================================================\n",
      "Epoch 2, Itrs 560/10000, Loss=0.907405\n",
      "Epoch 2, Itrs 570/10000, Loss=0.938491\n",
      "Epoch 2, Itrs 580/10000, Loss=0.959184\n",
      "Epoch 2, Itrs 590/10000, Loss=0.909472\n",
      "Epoch 2, Itrs 600/10000, Loss=0.944159\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 17.03it/s]\n",
      "Validation Loss: 0.478526\n",
      "\n",
      "Foreground Acc: 0.943417\n",
      "Precision: 0.911214\n",
      "Recall: 0.943417\n",
      "F1 Score: 0.927036\n",
      "IoU Foreground: 0.863996\n",
      "IoU Background: 0.961124\n",
      "Mean IoU: 0.912560\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600]\n",
      "==========================================================\n",
      "Epoch 2, Itrs 610/10000, Loss=0.887004\n",
      "Epoch 2, Itrs 620/10000, Loss=0.982831\n",
      "Epoch 2, Itrs 630/10000, Loss=0.985055\n",
      "Epoch 2, Itrs 640/10000, Loss=0.890362\n",
      "Epoch 2, Itrs 650/10000, Loss=0.950359\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.31it/s]\n",
      "Validation Loss: 0.484856\n",
      "\n",
      "Foreground Acc: 0.963906\n",
      "Precision: 0.887300\n",
      "Recall: 0.963906\n",
      "F1 Score: 0.924018\n",
      "IoU Foreground: 0.858768\n",
      "IoU Background: 0.958277\n",
      "Mean IoU: 0.908522\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650]\n",
      "==========================================================\n",
      "Epoch 2, Itrs 660/10000, Loss=0.893550\n",
      "Epoch 2, Itrs 670/10000, Loss=0.930216\n",
      "Epoch 2, Itrs 680/10000, Loss=0.892657\n",
      "Epoch 2, Itrs 690/10000, Loss=0.936214\n",
      "Epoch 2, Itrs 700/10000, Loss=0.950970\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.67it/s]\n",
      "Validation Loss: 0.563466\n",
      "\n",
      "Foreground Acc: 0.974078\n",
      "Precision: 0.861400\n",
      "Recall: 0.974078\n",
      "F1 Score: 0.914280\n",
      "IoU Foreground: 0.842096\n",
      "IoU Background: 0.951798\n",
      "Mean IoU: 0.896947\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797, 0.8420961102639317]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700]\n",
      "==========================================================\n",
      "Epoch 2, Itrs 710/10000, Loss=0.907301\n",
      "Epoch 2, Itrs 720/10000, Loss=0.918125\n",
      "Epoch 2, Itrs 730/10000, Loss=0.970979\n",
      "Epoch 2, Itrs 740/10000, Loss=0.929009\n",
      "Epoch 2, Itrs 750/10000, Loss=0.920783\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.43it/s]\n",
      "Validation Loss: 0.522022\n",
      "\n",
      "Foreground Acc: 0.924957\n",
      "Precision: 0.893217\n",
      "Recall: 0.924957\n",
      "F1 Score: 0.908810\n",
      "IoU Foreground: 0.832861\n",
      "IoU Background: 0.951641\n",
      "Mean IoU: 0.892251\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797, 0.8420961102639317, 0.8328608153562526]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750]\n",
      "==========================================================\n",
      "Epoch 2, Itrs 760/10000, Loss=0.951379\n",
      "Epoch 2, Itrs 770/10000, Loss=0.920402\n",
      "Epoch 2, Itrs 780/10000, Loss=1.002730\n",
      "Epoch 2, Itrs 790/10000, Loss=0.934010\n",
      "Epoch 2, Itrs 800/10000, Loss=0.909064\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 15.85it/s]\n",
      "Validation Loss: 0.468338\n",
      "\n",
      "Foreground Acc: 0.956071\n",
      "Precision: 0.901665\n",
      "Recall: 0.956071\n",
      "F1 Score: 0.928072\n",
      "IoU Foreground: 0.865796\n",
      "IoU Background: 0.961076\n",
      "Mean IoU: 0.913436\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797, 0.8420961102639317, 0.8328608153562526, 0.8657961442428483]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800]\n",
      "==========================================================\n",
      "Epoch 2, Itrs 810/10000, Loss=0.901209\n",
      "Epoch 2, Itrs 820/10000, Loss=0.883588\n",
      "Epoch 2, Itrs 830/10000, Loss=0.902844\n",
      "Epoch 2, Itrs 840/10000, Loss=0.935608\n",
      "Epoch 2, Itrs 850/10000, Loss=0.915038\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 15.64it/s]\n",
      "Validation Loss: 0.722810\n",
      "\n",
      "Foreground Acc: 0.984001\n",
      "Precision: 0.641993\n",
      "Recall: 0.984001\n",
      "F1 Score: 0.777028\n",
      "IoU Foreground: 0.635361\n",
      "IoU Background: 0.850576\n",
      "Mean IoU: 0.742968\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797, 0.8420961102639317, 0.8328608153562526, 0.8657961442428483, 0.6353605500721456]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850]\n",
      "==========================================================\n",
      "Epoch 2, Itrs 860/10000, Loss=0.987233\n",
      "Epoch 2, Itrs 870/10000, Loss=0.932299\n",
      "Epoch 2, Itrs 880/10000, Loss=0.905332\n",
      "Epoch 2, Itrs 890/10000, Loss=0.878758\n",
      "Epoch 2, Itrs 900/10000, Loss=0.951897\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 15.71it/s]\n",
      "Validation Loss: 0.583698\n",
      "\n",
      "Foreground Acc: 0.968362\n",
      "Precision: 0.820457\n",
      "Recall: 0.968362\n",
      "F1 Score: 0.888295\n",
      "IoU Foreground: 0.799038\n",
      "IoU Background: 0.935824\n",
      "Mean IoU: 0.867431\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797, 0.8420961102639317, 0.8328608153562526, 0.8657961442428483, 0.6353605500721456, 0.7990384840357662]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900]\n",
      "==========================================================\n",
      "Epoch 2, Itrs 910/10000, Loss=0.902940\n",
      "Epoch 2, Itrs 920/10000, Loss=0.936716\n",
      "Epoch 2, Itrs 930/10000, Loss=0.962199\n",
      "Epoch 3, Itrs 940/10000, Loss=0.895048\n",
      "Epoch 3, Itrs 950/10000, Loss=0.872068\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 15.89it/s]\n",
      "Validation Loss: 0.533433\n",
      "\n",
      "Foreground Acc: 0.975521\n",
      "Precision: 0.829550\n",
      "Recall: 0.975521\n",
      "F1 Score: 0.896633\n",
      "IoU Foreground: 0.812634\n",
      "IoU Background: 0.940620\n",
      "Mean IoU: 0.876627\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797, 0.8420961102639317, 0.8328608153562526, 0.8657961442428483, 0.6353605500721456, 0.7990384840357662, 0.8126335547561492]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950]\n",
      "==========================================================\n",
      "Epoch 3, Itrs 960/10000, Loss=0.931487\n",
      "Epoch 3, Itrs 970/10000, Loss=0.901159\n",
      "Epoch 3, Itrs 980/10000, Loss=0.866859\n",
      "Epoch 3, Itrs 990/10000, Loss=0.901571\n",
      "Epoch 3, Itrs 1000/10000, Loss=0.952213\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 15.93it/s]\n",
      "Validation Loss: 0.467007\n",
      "\n",
      "Foreground Acc: 0.954569\n",
      "Precision: 0.901169\n",
      "Recall: 0.954569\n",
      "F1 Score: 0.927101\n",
      "IoU Foreground: 0.864108\n",
      "IoU Background: 0.960586\n",
      "Mean IoU: 0.912347\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797, 0.8420961102639317, 0.8328608153562526, 0.8657961442428483, 0.6353605500721456, 0.7990384840357662, 0.8126335547561492, 0.8641076801266825]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000]\n",
      "==========================================================\n",
      "Epoch 3, Itrs 1010/10000, Loss=0.918595\n",
      "Epoch 3, Itrs 1020/10000, Loss=0.872846\n",
      "Epoch 3, Itrs 1030/10000, Loss=0.933315\n",
      "Epoch 3, Itrs 1040/10000, Loss=0.953829\n",
      "Epoch 3, Itrs 1050/10000, Loss=0.886286\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 17.08it/s]\n",
      "Validation Loss: 0.460675\n",
      "\n",
      "Foreground Acc: 0.972211\n",
      "Precision: 0.880964\n",
      "Recall: 0.972211\n",
      "F1 Score: 0.924341\n",
      "IoU Foreground: 0.859326\n",
      "IoU Background: 0.958020\n",
      "Mean IoU: 0.908673\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797, 0.8420961102639317, 0.8328608153562526, 0.8657961442428483, 0.6353605500721456, 0.7990384840357662, 0.8126335547561492, 0.8641076801266825, 0.8593256088694903]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050]\n",
      "==========================================================\n",
      "Epoch 3, Itrs 1060/10000, Loss=0.895918\n",
      "Epoch 3, Itrs 1070/10000, Loss=0.896821\n",
      "Epoch 3, Itrs 1080/10000, Loss=0.902769\n",
      "Epoch 3, Itrs 1090/10000, Loss=0.845012\n",
      "Epoch 3, Itrs 1100/10000, Loss=0.877978\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.28it/s]\n",
      "Validation Loss: 0.431240\n",
      "\n",
      "Foreground Acc: 0.913285\n",
      "Precision: 0.921565\n",
      "Recall: 0.913285\n",
      "F1 Score: 0.917406\n",
      "IoU Foreground: 0.847415\n",
      "IoU Background: 0.957288\n",
      "Mean IoU: 0.902351\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797, 0.8420961102639317, 0.8328608153562526, 0.8657961442428483, 0.6353605500721456, 0.7990384840357662, 0.8126335547561492, 0.8641076801266825, 0.8593256088694903, 0.8474151336663303]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100]\n",
      "==========================================================\n",
      "Epoch 3, Itrs 1110/10000, Loss=0.850436\n",
      "Epoch 3, Itrs 1120/10000, Loss=0.886972\n",
      "Epoch 3, Itrs 1130/10000, Loss=0.908566\n",
      "Epoch 3, Itrs 1140/10000, Loss=0.911821\n",
      "Epoch 3, Itrs 1150/10000, Loss=0.927288\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.46it/s]\n",
      "Validation Loss: 0.435605\n",
      "\n",
      "Foreground Acc: 0.960666\n",
      "Precision: 0.905351\n",
      "Recall: 0.960666\n",
      "F1 Score: 0.932188\n",
      "IoU Foreground: 0.872990\n",
      "IoU Background: 0.963246\n",
      "Mean IoU: 0.918118\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797, 0.8420961102639317, 0.8328608153562526, 0.8657961442428483, 0.6353605500721456, 0.7990384840357662, 0.8126335547561492, 0.8641076801266825, 0.8593256088694903, 0.8474151336663303, 0.8729896976681869]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150]\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_without_prob0.8729896976681869.pth\n",
      "saved here.\n",
      "==========================================================\n",
      "Epoch 3, Itrs 1160/10000, Loss=0.959871\n",
      "Epoch 3, Itrs 1170/10000, Loss=0.853645\n",
      "Epoch 3, Itrs 1180/10000, Loss=0.956190\n",
      "Epoch 3, Itrs 1190/10000, Loss=0.989076\n",
      "Epoch 3, Itrs 1200/10000, Loss=0.940753\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.58it/s]\n",
      "Validation Loss: 0.496499\n",
      "\n",
      "Foreground Acc: 0.971589\n",
      "Precision: 0.885274\n",
      "Recall: 0.971589\n",
      "F1 Score: 0.926425\n",
      "IoU Foreground: 0.862935\n",
      "IoU Background: 0.959301\n",
      "Mean IoU: 0.911118\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797, 0.8420961102639317, 0.8328608153562526, 0.8657961442428483, 0.6353605500721456, 0.7990384840357662, 0.8126335547561492, 0.8641076801266825, 0.8593256088694903, 0.8474151336663303, 0.8729896976681869, 0.8629352151989066]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200]\n",
      "==========================================================\n",
      "Epoch 3, Itrs 1210/10000, Loss=0.997315\n",
      "Epoch 3, Itrs 1220/10000, Loss=0.881693\n",
      "Epoch 3, Itrs 1230/10000, Loss=0.932909\n",
      "Epoch 3, Itrs 1240/10000, Loss=0.927577\n",
      "Epoch 3, Itrs 1250/10000, Loss=0.918781\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.67it/s]\n",
      "Validation Loss: 0.480405\n",
      "\n",
      "Foreground Acc: 0.961991\n",
      "Precision: 0.892443\n",
      "Recall: 0.961991\n",
      "F1 Score: 0.925913\n",
      "IoU Foreground: 0.862046\n",
      "IoU Background: 0.959502\n",
      "Mean IoU: 0.910774\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797, 0.8420961102639317, 0.8328608153562526, 0.8657961442428483, 0.6353605500721456, 0.7990384840357662, 0.8126335547561492, 0.8641076801266825, 0.8593256088694903, 0.8474151336663303, 0.8729896976681869, 0.8629352151989066, 0.8620463568021361]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250]\n",
      "==========================================================\n",
      "Epoch 3, Itrs 1260/10000, Loss=0.906047\n",
      "Epoch 3, Itrs 1270/10000, Loss=0.894572\n",
      "Epoch 3, Itrs 1280/10000, Loss=0.915070\n",
      "Epoch 3, Itrs 1290/10000, Loss=0.894545\n",
      "Epoch 3, Itrs 1300/10000, Loss=0.940683\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.68it/s]\n",
      "Validation Loss: 0.489606\n",
      "\n",
      "Foreground Acc: 0.960191\n",
      "Precision: 0.891921\n",
      "Recall: 0.960191\n",
      "F1 Score: 0.924798\n",
      "IoU Foreground: 0.860116\n",
      "IoU Background: 0.958940\n",
      "Mean IoU: 0.909528\n",
      "\n",
      "[0.6825586295035073, 0.7724098165632423, 0.7797975776744844, 0.7812819754598834, 0.7709820873395267, 0.8450568952991855, 0.6739351654031348, 0.81676808852988, 0.8047349783257012, 0.8504620350701382, 0.8700079479483994, 0.86399560540522, 0.8587675120823797, 0.8420961102639317, 0.8328608153562526, 0.8657961442428483, 0.6353605500721456, 0.7990384840357662, 0.8126335547561492, 0.8641076801266825, 0.8593256088694903, 0.8474151336663303, 0.8729896976681869, 0.8629352151989066, 0.8620463568021361, 0.8601158986047025]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300]\n",
      "==========================================================\n",
      "Epoch 3, Itrs 1310/10000, Loss=0.886003\n",
      "^C\n",
      "Traceback (most recent call last):\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/leaf_train.py\", line 536, in <module>\n",
      "    main()\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/leaf_train.py\", line 463, in main\n",
      "    np_loss = loss.detach().cpu().numpy()\n",
      "KeyboardInterrupt\n"
     ]
    }
   ],
   "source": [
    "!python3 leaf_train.py --model deeplabv3plus_mobilenet --custom_data_path \"crf_sam_annotations_large/\" --batch_size 4 --lr 0.12 --total_itrs 10000 --loss_type binary_dice --val_interval 50  --separable_conv   \n",
    "# /teamspace/studios/this_studio/Deeplab/saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8213798905479425.pth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "ff19a918-de34-40d6-a833-21f156d0a460",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device: cuda\n",
      "Dataset: custom, Train set: 1860, Val set: 496\n",
      "[!] Retrain\n",
      "Epoch 1, Itrs 10/10000, Loss=1.762195\n",
      "Epoch 1, Itrs 20/10000, Loss=1.600969\n",
      "Epoch 1, Itrs 30/10000, Loss=1.633991\n",
      "Epoch 1, Itrs 40/10000, Loss=1.548969\n",
      "Epoch 1, Itrs 50/10000, Loss=1.331846\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.38it/s]\n",
      "Validation Loss: 0.729322\n",
      "\n",
      "Foreground Acc: 0.889402\n",
      "Precision: 0.716121\n",
      "Recall: 0.889402\n",
      "F1 Score: 0.793410\n",
      "IoU Foreground: 0.657564\n",
      "IoU Background: 0.880441\n",
      "Mean IoU: 0.769003\n",
      "\n",
      "[0.6575644882132871]\n",
      "[50]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 1, Itrs 60/10000, Loss=1.499390\n",
      "Epoch 1, Itrs 70/10000, Loss=1.454118\n",
      "Epoch 1, Itrs 80/10000, Loss=1.457877\n",
      "Epoch 1, Itrs 90/10000, Loss=1.348878\n",
      "Epoch 1, Itrs 100/10000, Loss=1.404372\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.47it/s]\n",
      "Validation Loss: 0.403181\n",
      "\n",
      "Foreground Acc: 0.917555\n",
      "Precision: 0.846008\n",
      "Recall: 0.917555\n",
      "F1 Score: 0.880330\n",
      "IoU Foreground: 0.786241\n",
      "IoU Background: 0.935135\n",
      "Mean IoU: 0.860688\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815]\n",
      "[50, 100]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 1, Itrs 110/10000, Loss=1.285143\n",
      "Epoch 1, Itrs 120/10000, Loss=1.351703\n",
      "Epoch 1, Itrs 130/10000, Loss=1.282665\n",
      "Epoch 1, Itrs 140/10000, Loss=1.416689\n",
      "Epoch 1, Itrs 150/10000, Loss=1.295390\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.62it/s]\n",
      "Validation Loss: 0.491468\n",
      "\n",
      "Foreground Acc: 0.837282\n",
      "Precision: 0.923977\n",
      "Recall: 0.837282\n",
      "F1 Score: 0.878496\n",
      "IoU Foreground: 0.783319\n",
      "IoU Background: 0.941008\n",
      "Mean IoU: 0.862163\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338]\n",
      "[50, 100, 150]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 1, Itrs 160/10000, Loss=1.378120\n",
      "Epoch 1, Itrs 170/10000, Loss=1.434958\n",
      "Epoch 1, Itrs 180/10000, Loss=1.320741\n",
      "Epoch 1, Itrs 190/10000, Loss=1.327690\n",
      "Epoch 1, Itrs 200/10000, Loss=1.287202\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.56it/s]\n",
      "Validation Loss: 0.526494\n",
      "\n",
      "Foreground Acc: 0.926868\n",
      "Precision: 0.883296\n",
      "Recall: 0.926868\n",
      "F1 Score: 0.904558\n",
      "IoU Foreground: 0.825746\n",
      "IoU Background: 0.949018\n",
      "Mean IoU: 0.887382\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819]\n",
      "[50, 100, 150, 200]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 1, Itrs 210/10000, Loss=1.304009\n",
      "Epoch 1, Itrs 220/10000, Loss=1.400336\n",
      "Epoch 1, Itrs 230/10000, Loss=1.260629\n",
      "Epoch 1, Itrs 240/10000, Loss=1.291886\n",
      "Epoch 1, Itrs 250/10000, Loss=1.235138\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.33it/s]\n",
      "Validation Loss: 0.447416\n",
      "\n",
      "Foreground Acc: 0.939262\n",
      "Precision: 0.898355\n",
      "Recall: 0.939262\n",
      "F1 Score: 0.918353\n",
      "IoU Foreground: 0.849033\n",
      "IoU Background: 0.956327\n",
      "Mean IoU: 0.902680\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221]\n",
      "[50, 100, 150, 200, 250]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 1, Itrs 260/10000, Loss=1.248781\n",
      "Epoch 1, Itrs 270/10000, Loss=1.231850\n",
      "Epoch 1, Itrs 280/10000, Loss=1.452824\n",
      "Epoch 1, Itrs 290/10000, Loss=1.202333\n",
      "Epoch 1, Itrs 300/10000, Loss=1.294753\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.21it/s]\n",
      "Validation Loss: 0.371241\n",
      "\n",
      "Foreground Acc: 0.914179\n",
      "Precision: 0.916187\n",
      "Recall: 0.914179\n",
      "F1 Score: 0.915182\n",
      "IoU Foreground: 0.843627\n",
      "IoU Background: 0.955978\n",
      "Mean IoU: 0.899802\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166]\n",
      "[50, 100, 150, 200, 250, 300]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 1, Itrs 310/10000, Loss=1.312380\n",
      "Epoch 1, Itrs 320/10000, Loss=1.415913\n",
      "Epoch 1, Itrs 330/10000, Loss=1.432448\n",
      "Epoch 1, Itrs 340/10000, Loss=1.379366\n",
      "Epoch 1, Itrs 350/10000, Loss=1.143553\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.31it/s]\n",
      "Validation Loss: 0.520406\n",
      "\n",
      "Foreground Acc: 0.969308\n",
      "Precision: 0.831953\n",
      "Recall: 0.969308\n",
      "F1 Score: 0.895393\n",
      "IoU Foreground: 0.810599\n",
      "IoU Background: 0.940306\n",
      "Mean IoU: 0.875452\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344]\n",
      "[50, 100, 150, 200, 250, 300, 350]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 1, Itrs 360/10000, Loss=1.125245\n",
      "Epoch 1, Itrs 370/10000, Loss=1.325924\n",
      "Epoch 1, Itrs 380/10000, Loss=1.234267\n",
      "Epoch 1, Itrs 390/10000, Loss=1.355711\n",
      "Epoch 1, Itrs 400/10000, Loss=1.315139\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.27it/s]\n",
      "Validation Loss: 0.370838\n",
      "\n",
      "Foreground Acc: 0.955716\n",
      "Precision: 0.876731\n",
      "Recall: 0.955716\n",
      "F1 Score: 0.914521\n",
      "IoU Foreground: 0.842504\n",
      "IoU Background: 0.953079\n",
      "Mean IoU: 0.897792\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344, 0.8425044455785039]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 1, Itrs 410/10000, Loss=1.216761\n",
      "Epoch 1, Itrs 420/10000, Loss=1.395643\n",
      "Epoch 1, Itrs 430/10000, Loss=1.240946\n",
      "Epoch 1, Itrs 440/10000, Loss=1.260945\n",
      "Epoch 1, Itrs 450/10000, Loss=1.285742\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.74it/s]\n",
      "Validation Loss: 0.431866\n",
      "\n",
      "Foreground Acc: 0.956174\n",
      "Precision: 0.901239\n",
      "Recall: 0.956174\n",
      "F1 Score: 0.927894\n",
      "IoU Foreground: 0.865487\n",
      "IoU Background: 0.960967\n",
      "Mean IoU: 0.913227\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344, 0.8425044455785039, 0.8654870011570238]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 1, Itrs 460/10000, Loss=1.424003\n",
      "Epoch 2, Itrs 470/10000, Loss=1.280303\n",
      "Epoch 2, Itrs 480/10000, Loss=1.365350\n",
      "Epoch 2, Itrs 490/10000, Loss=1.292265\n",
      "Epoch 2, Itrs 500/10000, Loss=1.350322\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.39it/s]\n",
      "Validation Loss: 0.470549\n",
      "\n",
      "Foreground Acc: 0.970352\n",
      "Precision: 0.860413\n",
      "Recall: 0.970352\n",
      "F1 Score: 0.912082\n",
      "IoU Foreground: 0.838373\n",
      "IoU Background: 0.950680\n",
      "Mean IoU: 0.894527\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344, 0.8425044455785039, 0.8654870011570238, 0.8383733241745189]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 2, Itrs 510/10000, Loss=1.178868\n",
      "Epoch 2, Itrs 520/10000, Loss=1.242084\n",
      "Epoch 2, Itrs 530/10000, Loss=1.193247\n",
      "Epoch 2, Itrs 540/10000, Loss=1.069832\n",
      "Epoch 2, Itrs 550/10000, Loss=1.231092\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.29it/s]\n",
      "Validation Loss: 0.447172\n",
      "\n",
      "Foreground Acc: 0.965419\n",
      "Precision: 0.894100\n",
      "Recall: 0.965419\n",
      "F1 Score: 0.928392\n",
      "IoU Foreground: 0.866354\n",
      "IoU Background: 0.960787\n",
      "Mean IoU: 0.913570\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344, 0.8425044455785039, 0.8654870011570238, 0.8383733241745189, 0.8663536283440466]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 2, Itrs 560/10000, Loss=1.187540\n",
      "Epoch 2, Itrs 570/10000, Loss=1.250801\n",
      "Epoch 2, Itrs 580/10000, Loss=1.327641\n",
      "Epoch 2, Itrs 590/10000, Loss=1.262100\n",
      "Epoch 2, Itrs 600/10000, Loss=1.246909\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.72it/s]\n",
      "Validation Loss: 0.400199\n",
      "\n",
      "Foreground Acc: 0.954572\n",
      "Precision: 0.908366\n",
      "Recall: 0.954572\n",
      "F1 Score: 0.930896\n",
      "IoU Foreground: 0.870725\n",
      "IoU Background: 0.962790\n",
      "Mean IoU: 0.916758\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344, 0.8425044455785039, 0.8654870011570238, 0.8383733241745189, 0.8663536283440466, 0.8707248307045039]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 2, Itrs 610/10000, Loss=1.161459\n",
      "Epoch 2, Itrs 620/10000, Loss=1.372982\n",
      "Epoch 2, Itrs 630/10000, Loss=1.395354\n",
      "Epoch 2, Itrs 640/10000, Loss=1.232112\n",
      "Epoch 2, Itrs 650/10000, Loss=1.282644\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.74it/s]\n",
      "Validation Loss: 0.452413\n",
      "\n",
      "Foreground Acc: 0.970194\n",
      "Precision: 0.880766\n",
      "Recall: 0.970194\n",
      "F1 Score: 0.923320\n",
      "IoU Foreground: 0.857562\n",
      "IoU Background: 0.957517\n",
      "Mean IoU: 0.907539\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344, 0.8425044455785039, 0.8654870011570238, 0.8383733241745189, 0.8663536283440466, 0.8707248307045039, 0.8575615200328687]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 2, Itrs 660/10000, Loss=1.223464\n",
      "Epoch 2, Itrs 670/10000, Loss=1.277164\n",
      "Epoch 2, Itrs 680/10000, Loss=1.206791\n",
      "Epoch 2, Itrs 690/10000, Loss=1.292240\n",
      "Epoch 2, Itrs 700/10000, Loss=1.281439\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.63it/s]\n",
      "Validation Loss: 0.464227\n",
      "\n",
      "Foreground Acc: 0.976631\n",
      "Precision: 0.853643\n",
      "Recall: 0.976631\n",
      "F1 Score: 0.911005\n",
      "IoU Foreground: 0.836555\n",
      "IoU Background: 0.949610\n",
      "Mean IoU: 0.893083\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344, 0.8425044455785039, 0.8654870011570238, 0.8383733241745189, 0.8663536283440466, 0.8707248307045039, 0.8575615200328687, 0.8365553881870408]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 2, Itrs 710/10000, Loss=1.199645\n",
      "Epoch 2, Itrs 720/10000, Loss=1.218055\n",
      "Epoch 2, Itrs 730/10000, Loss=1.321610\n",
      "Epoch 2, Itrs 740/10000, Loss=1.276626\n",
      "Epoch 2, Itrs 750/10000, Loss=1.186007\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.67it/s]\n",
      "Validation Loss: 0.352272\n",
      "\n",
      "Foreground Acc: 0.966109\n",
      "Precision: 0.877777\n",
      "Recall: 0.966109\n",
      "F1 Score: 0.919827\n",
      "IoU Foreground: 0.851556\n",
      "IoU Background: 0.955649\n",
      "Mean IoU: 0.903602\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344, 0.8425044455785039, 0.8654870011570238, 0.8383733241745189, 0.8663536283440466, 0.8707248307045039, 0.8575615200328687, 0.8365553881870408, 0.8515560733577786]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 2, Itrs 760/10000, Loss=1.295136\n",
      "Epoch 2, Itrs 770/10000, Loss=1.203767\n",
      "Epoch 2, Itrs 780/10000, Loss=1.479200\n",
      "Epoch 2, Itrs 790/10000, Loss=1.302796\n",
      "Epoch 2, Itrs 800/10000, Loss=1.203925\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.89it/s]\n",
      "Validation Loss: 0.426963\n",
      "\n",
      "Foreground Acc: 0.974097\n",
      "Precision: 0.866444\n",
      "Recall: 0.974097\n",
      "F1 Score: 0.917122\n",
      "IoU Foreground: 0.846931\n",
      "IoU Background: 0.953539\n",
      "Mean IoU: 0.900235\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344, 0.8425044455785039, 0.8654870011570238, 0.8383733241745189, 0.8663536283440466, 0.8707248307045039, 0.8575615200328687, 0.8365553881870408, 0.8515560733577786, 0.8469307393463604]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 2, Itrs 810/10000, Loss=1.188221\n",
      "Epoch 2, Itrs 820/10000, Loss=1.180668\n",
      "Epoch 2, Itrs 830/10000, Loss=1.229000\n",
      "Epoch 2, Itrs 840/10000, Loss=1.247999\n",
      "Epoch 2, Itrs 850/10000, Loss=1.199250\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.91it/s]\n",
      "Validation Loss: 0.416585\n",
      "\n",
      "Foreground Acc: 0.959332\n",
      "Precision: 0.879173\n",
      "Recall: 0.959332\n",
      "F1 Score: 0.917505\n",
      "IoU Foreground: 0.847583\n",
      "IoU Background: 0.954650\n",
      "Mean IoU: 0.901117\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344, 0.8425044455785039, 0.8654870011570238, 0.8383733241745189, 0.8663536283440466, 0.8707248307045039, 0.8575615200328687, 0.8365553881870408, 0.8515560733577786, 0.8469307393463604, 0.8475833153586957]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 2, Itrs 860/10000, Loss=1.377831\n",
      "Epoch 2, Itrs 870/10000, Loss=1.253586\n",
      "Epoch 2, Itrs 880/10000, Loss=1.130307\n",
      "Epoch 2, Itrs 890/10000, Loss=1.212267\n",
      "Epoch 2, Itrs 900/10000, Loss=1.263062\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.64it/s]\n",
      "Validation Loss: 0.389410\n",
      "\n",
      "Foreground Acc: 0.968828\n",
      "Precision: 0.899123\n",
      "Recall: 0.968828\n",
      "F1 Score: 0.932675\n",
      "IoU Foreground: 0.873843\n",
      "IoU Background: 0.963139\n",
      "Mean IoU: 0.918491\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344, 0.8425044455785039, 0.8654870011570238, 0.8383733241745189, 0.8663536283440466, 0.8707248307045039, 0.8575615200328687, 0.8365553881870408, 0.8515560733577786, 0.8469307393463604, 0.8475833153586957, 0.8738434159526594]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 2, Itrs 910/10000, Loss=1.160888\n",
      "Epoch 2, Itrs 920/10000, Loss=1.265235\n",
      "Epoch 2, Itrs 930/10000, Loss=1.292252\n",
      "Epoch 3, Itrs 940/10000, Loss=1.149728\n",
      "Epoch 3, Itrs 950/10000, Loss=1.146365\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.95it/s]\n",
      "Validation Loss: 0.398922\n",
      "\n",
      "Foreground Acc: 0.972409\n",
      "Precision: 0.870742\n",
      "Recall: 0.972409\n",
      "F1 Score: 0.918772\n",
      "IoU Foreground: 0.849748\n",
      "IoU Background: 0.954645\n",
      "Mean IoU: 0.902196\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344, 0.8425044455785039, 0.8654870011570238, 0.8383733241745189, 0.8663536283440466, 0.8707248307045039, 0.8575615200328687, 0.8365553881870408, 0.8515560733577786, 0.8469307393463604, 0.8475833153586957, 0.8738434159526594, 0.8497480917359952]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 3, Itrs 960/10000, Loss=1.211539\n",
      "Epoch 3, Itrs 970/10000, Loss=1.231361\n",
      "Epoch 3, Itrs 980/10000, Loss=1.149149\n",
      "Epoch 3, Itrs 990/10000, Loss=1.211500\n",
      "Epoch 3, Itrs 1000/10000, Loss=1.325231\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.96it/s]\n",
      "Validation Loss: 0.368966\n",
      "\n",
      "Foreground Acc: 0.957317\n",
      "Precision: 0.928649\n",
      "Recall: 0.957317\n",
      "F1 Score: 0.942765\n",
      "IoU Foreground: 0.891727\n",
      "IoU Background: 0.969460\n",
      "Mean IoU: 0.930594\n",
      "\n",
      "[0.6575644882132871, 0.7862411413462815, 0.783318896723338, 0.8257464316632819, 0.8490326407078221, 0.84362701594166, 0.8105991474634344, 0.8425044455785039, 0.8654870011570238, 0.8383733241745189, 0.8663536283440466, 0.8707248307045039, 0.8575615200328687, 0.8365553881870408, 0.8515560733577786, 0.8469307393463604, 0.8475833153586957, 0.8738434159526594, 0.8497480917359952, 0.8917271802772625]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 3, Itrs 1010/10000, Loss=1.284667\n",
      "Epoch 3, Itrs 1020/10000, Loss=1.185542\n",
      "Epoch 3, Itrs 1030/10000, Loss=1.262347\n",
      "Epoch 3, Itrs 1040/10000, Loss=1.284819\n",
      "Epoch 3, Itrs 1050/10000, Loss=1.171552\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "39it [00:02, 18.00it/s]^C\n",
      "39it [00:02, 16.62it/s]\n",
      "Traceback (most recent call last):\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/binary_leaf_kl_updated.py\", line 647, in <module>\n",
      "    main()\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/binary_leaf_kl_updated.py\", line 589, in main\n",
      "    val_score, ret_samples, current_val_loss = validate(\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/binary_leaf_kl_updated.py\", line 339, in validate\n",
      "    targets = labels.cpu().numpy()\n",
      "KeyboardInterrupt\n"
     ]
    }
   ],
   "source": [
    "!python3 binary_leaf_kl_updated.py --model deeplabv3plus_mobilenet --custom_data_path \"crf_sam_annotations_large/\" --batch_size 4 --lr 0.12 --total_itrs 10000 --loss_type kl_dice_combined --val_interval 50  --separable_conv   \n",
    "# /teamspace/studios/this_studio/Deeplab/saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8213798905479425.pth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "cb5eb83a-40a6-48df-8842-5920718878d3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device: cuda\n",
      "Dataset: custom, Train set: 186, Val set: 50\n",
      "[!] Retrain\n",
      "Epoch 1, Itrs 10/10000, Loss=1.219968\n",
      "Epoch 1, Itrs 20/10000, Loss=1.211644\n",
      "Epoch 1, Itrs 30/10000, Loss=1.150467\n",
      "Epoch 1, Itrs 40/10000, Loss=1.040683\n",
      "Epoch 2, Itrs 50/10000, Loss=1.213698\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "13it [00:00, 14.20it/s]\n",
      "Validation Loss: 0.415333\n",
      "\n",
      "Foreground Acc: 0.929049\n",
      "Precision: 0.832942\n",
      "Recall: 0.929049\n",
      "F1 Score: 0.878374\n",
      "IoU Foreground: 0.783126\n",
      "IoU Background: 0.921587\n",
      "Mean IoU: 0.852357\n",
      "\n",
      "[0.7831262504105422]\n",
      "[50]\n",
      "==========================================================\n",
      "Epoch 2, Itrs 60/10000, Loss=1.220376\n",
      "^C\n",
      "Traceback (most recent call last):\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/binary_leaf_main.py\", line 596, in <module>\n",
      "    main()\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/binary_leaf_main.py\", line 503, in main\n",
      "    labels = labels.to(device, dtype=torch.long)\n",
      "KeyboardInterrupt\n"
     ]
    }
   ],
   "source": [
    "!python3 binary_leaf_main.py --model deeplabv3plus_mobilenet --custom_data_path \"crf_sam_annotations_large/\" --batch_size 4 --lr 0.12 --total_itrs 10000 --loss_type binary_dice --val_interval 50  --separable_conv   \n",
    "# /teamspace/studios/this_studio/Deeplab/saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8213798905479425.pth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "97bc6881-81da-4188-963a-6ce58d480965",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device: cuda\n",
      "Dataset: custom, Train set: 1860, Val set: 496\n",
      "[!] Retrain\n",
      "Epoch 1, Itrs 10/10000, Loss=1.761071\n",
      "Epoch 1, Itrs 20/10000, Loss=1.618937\n",
      "Epoch 1, Itrs 30/10000, Loss=1.670326\n",
      "Epoch 1, Itrs 40/10000, Loss=1.577848\n",
      "Epoch 1, Itrs 50/10000, Loss=1.338472\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.17it/s]\n",
      "Validation Loss: 0.796973\n",
      "\n",
      "Foreground Acc: 0.914325\n",
      "Precision: 0.679277\n",
      "Recall: 0.914325\n",
      "F1 Score: 0.779467\n",
      "IoU Foreground: 0.638628\n",
      "IoU Background: 0.865583\n",
      "Mean IoU: 0.752105\n",
      "\n",
      "[0.6386279069071205]\n",
      "[50]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_with_prob_0.6386279069071205.pth\n",
      "==========================================================\n",
      "Epoch 1, Itrs 60/10000, Loss=1.490427\n",
      "Epoch 1, Itrs 70/10000, Loss=1.459067\n",
      "Epoch 1, Itrs 80/10000, Loss=1.446378\n",
      "Epoch 1, Itrs 90/10000, Loss=1.329029\n",
      "Epoch 1, Itrs 100/10000, Loss=1.380384\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.53it/s]\n",
      "Validation Loss: 0.490233\n",
      "\n",
      "Foreground Acc: 0.919748\n",
      "Precision: 0.818941\n",
      "Recall: 0.919748\n",
      "F1 Score: 0.866422\n",
      "IoU Foreground: 0.764325\n",
      "IoU Background: 0.926216\n",
      "Mean IoU: 0.845271\n",
      "\n",
      "[0.6386279069071205, 0.7643254213630535]\n",
      "[50, 100]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_with_prob_0.7643254213630535.pth\n",
      "==========================================================\n",
      "Epoch 1, Itrs 110/10000, Loss=1.281213\n",
      "Epoch 1, Itrs 120/10000, Loss=1.371097\n",
      "Epoch 1, Itrs 130/10000, Loss=1.300144\n",
      "Epoch 1, Itrs 140/10000, Loss=1.412707\n",
      "Epoch 1, Itrs 150/10000, Loss=1.301357\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.56it/s]\n",
      "Validation Loss: 0.534533\n",
      "\n",
      "Foreground Acc: 0.792390\n",
      "Precision: 0.931754\n",
      "Recall: 0.792390\n",
      "F1 Score: 0.856440\n",
      "IoU Foreground: 0.748924\n",
      "IoU Background: 0.933103\n",
      "Mean IoU: 0.841013\n",
      "\n",
      "[0.6386279069071205, 0.7643254213630535, 0.7489237904776817]\n",
      "[50, 100, 150]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 1, Itrs 160/10000, Loss=1.379979\n",
      "Epoch 1, Itrs 170/10000, Loss=1.419443\n",
      "Epoch 1, Itrs 180/10000, Loss=1.301228\n",
      "Epoch 1, Itrs 190/10000, Loss=1.321569\n",
      "Epoch 1, Itrs 200/10000, Loss=1.280515\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.49it/s]\n",
      "Validation Loss: 0.448416\n",
      "\n",
      "Foreground Acc: 0.946273\n",
      "Precision: 0.872233\n",
      "Recall: 0.946273\n",
      "F1 Score: 0.907746\n",
      "IoU Foreground: 0.831076\n",
      "IoU Background: 0.949611\n",
      "Mean IoU: 0.890343\n",
      "\n",
      "[0.6386279069071205, 0.7643254213630535, 0.7489237904776817, 0.831075799401631]\n",
      "[50, 100, 150, 200]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_with_prob_0.831075799401631.pth\n",
      "==========================================================\n",
      "Epoch 1, Itrs 210/10000, Loss=1.303319\n",
      "Epoch 1, Itrs 220/10000, Loss=1.402650\n",
      "Epoch 1, Itrs 230/10000, Loss=1.249416\n",
      "Epoch 1, Itrs 240/10000, Loss=1.290602\n",
      "Epoch 1, Itrs 250/10000, Loss=1.246291\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.30it/s]\n",
      "Validation Loss: 0.389177\n",
      "\n",
      "Foreground Acc: 0.928817\n",
      "Precision: 0.904546\n",
      "Recall: 0.928817\n",
      "F1 Score: 0.916521\n",
      "IoU Foreground: 0.845906\n",
      "IoU Background: 0.955875\n",
      "Mean IoU: 0.900891\n",
      "\n",
      "[0.6386279069071205, 0.7643254213630535, 0.7489237904776817, 0.831075799401631, 0.8459055886071709]\n",
      "[50, 100, 150, 200, 250]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_with_prob_0.8459055886071709.pth\n",
      "==========================================================\n",
      "Epoch 1, Itrs 260/10000, Loss=1.266353\n",
      "Epoch 1, Itrs 270/10000, Loss=1.239940\n",
      "Epoch 1, Itrs 280/10000, Loss=1.454346\n",
      "Epoch 1, Itrs 290/10000, Loss=1.215530\n",
      "Epoch 1, Itrs 300/10000, Loss=1.309240\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.17it/s]\n",
      "Validation Loss: 0.430651\n",
      "\n",
      "Foreground Acc: 0.929705\n",
      "Precision: 0.842613\n",
      "Recall: 0.929705\n",
      "F1 Score: 0.884019\n",
      "IoU Foreground: 0.792145\n",
      "IoU Background: 0.936366\n",
      "Mean IoU: 0.864256\n",
      "\n",
      "[0.6386279069071205, 0.7643254213630535, 0.7489237904776817, 0.831075799401631, 0.8459055886071709, 0.7921450544677175]\n",
      "[50, 100, 150, 200, 250, 300]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 1, Itrs 310/10000, Loss=1.310189\n",
      "Epoch 1, Itrs 320/10000, Loss=1.404935\n",
      "Epoch 1, Itrs 330/10000, Loss=1.449083\n",
      "Epoch 1, Itrs 340/10000, Loss=1.423033\n",
      "Epoch 1, Itrs 350/10000, Loss=1.169817\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.14it/s]\n",
      "Validation Loss: 0.524411\n",
      "\n",
      "Foreground Acc: 0.976788\n",
      "Precision: 0.783638\n",
      "Recall: 0.976788\n",
      "F1 Score: 0.869617\n",
      "IoU Foreground: 0.769312\n",
      "IoU Background: 0.922647\n",
      "Mean IoU: 0.845980\n",
      "\n",
      "[0.6386279069071205, 0.7643254213630535, 0.7489237904776817, 0.831075799401631, 0.8459055886071709, 0.7921450544677175, 0.7693117470468316]\n",
      "[50, 100, 150, 200, 250, 300, 350]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 1, Itrs 360/10000, Loss=1.143581\n",
      "Epoch 1, Itrs 370/10000, Loss=1.340595\n",
      "Epoch 1, Itrs 380/10000, Loss=1.251430\n",
      "Epoch 1, Itrs 390/10000, Loss=1.361377\n",
      "Epoch 1, Itrs 400/10000, Loss=1.324194\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.41it/s]\n",
      "Validation Loss: 0.314800\n",
      "\n",
      "Foreground Acc: 0.956911\n",
      "Precision: 0.882701\n",
      "Recall: 0.956911\n",
      "F1 Score: 0.918309\n",
      "IoU Foreground: 0.848957\n",
      "IoU Background: 0.955274\n",
      "Mean IoU: 0.902115\n",
      "\n",
      "[0.6386279069071205, 0.7643254213630535, 0.7489237904776817, 0.831075799401631, 0.8459055886071709, 0.7921450544677175, 0.7693117470468316, 0.8489570569614452]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_with_prob_0.8489570569614452.pth\n",
      "==========================================================\n",
      "Epoch 1, Itrs 410/10000, Loss=1.229000\n",
      "Epoch 1, Itrs 420/10000, Loss=1.398435\n",
      "Epoch 1, Itrs 430/10000, Loss=1.233570\n",
      "Epoch 1, Itrs 440/10000, Loss=1.257140\n",
      "Epoch 1, Itrs 450/10000, Loss=1.284012\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.37it/s]\n",
      "Validation Loss: 0.487181\n",
      "\n",
      "Foreground Acc: 0.955680\n",
      "Precision: 0.887943\n",
      "Recall: 0.955680\n",
      "F1 Score: 0.920568\n",
      "IoU Foreground: 0.852826\n",
      "IoU Background: 0.956687\n",
      "Mean IoU: 0.904756\n",
      "\n",
      "[0.6386279069071205, 0.7643254213630535, 0.7489237904776817, 0.831075799401631, 0.8459055886071709, 0.7921450544677175, 0.7693117470468316, 0.8489570569614452, 0.8528256192134294]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_with_prob_0.8528256192134294.pth\n",
      "==========================================================\n",
      "Epoch 1, Itrs 460/10000, Loss=1.436550\n",
      "Epoch 2, Itrs 470/10000, Loss=1.275506\n",
      "Epoch 2, Itrs 480/10000, Loss=1.362544\n",
      "Epoch 2, Itrs 490/10000, Loss=1.296493\n",
      "Epoch 2, Itrs 500/10000, Loss=1.345614\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.44it/s]\n",
      "Validation Loss: 0.401857\n",
      "\n",
      "Foreground Acc: 0.952841\n",
      "Precision: 0.895875\n",
      "Recall: 0.952841\n",
      "F1 Score: 0.923480\n",
      "IoU Foreground: 0.857839\n",
      "IoU Background: 0.958561\n",
      "Mean IoU: 0.908200\n",
      "\n",
      "[0.6386279069071205, 0.7643254213630535, 0.7489237904776817, 0.831075799401631, 0.8459055886071709, 0.7921450544677175, 0.7693117470468316, 0.8489570569614452, 0.8528256192134294, 0.8578387507996919]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_with_prob_0.8578387507996919.pth\n",
      "==========================================================\n",
      "Epoch 2, Itrs 510/10000, Loss=1.182999\n",
      "Epoch 2, Itrs 520/10000, Loss=1.250707\n",
      "Epoch 2, Itrs 530/10000, Loss=1.201606\n",
      "Epoch 2, Itrs 540/10000, Loss=1.068368\n",
      "Epoch 2, Itrs 550/10000, Loss=1.223182\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.73it/s]\n",
      "Validation Loss: 0.488042\n",
      "\n",
      "Foreground Acc: 0.955680\n",
      "Precision: 0.894763\n",
      "Recall: 0.955680\n",
      "F1 Score: 0.924219\n",
      "IoU Foreground: 0.859114\n",
      "IoU Background: 0.958841\n",
      "Mean IoU: 0.908977\n",
      "\n",
      "[0.6386279069071205, 0.7643254213630535, 0.7489237904776817, 0.831075799401631, 0.8459055886071709, 0.7921450544677175, 0.7693117470468316, 0.8489570569614452, 0.8528256192134294, 0.8578387507996919, 0.8591140226494551]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_with_prob_0.8591140226494551.pth\n",
      "==========================================================\n",
      "Epoch 2, Itrs 560/10000, Loss=1.186661\n",
      "Epoch 2, Itrs 570/10000, Loss=1.255578\n",
      "Epoch 2, Itrs 580/10000, Loss=1.326619\n",
      "Epoch 2, Itrs 590/10000, Loss=1.264220\n",
      "Epoch 2, Itrs 600/10000, Loss=1.243131\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.56it/s]\n",
      "Validation Loss: 0.392424\n",
      "\n",
      "Foreground Acc: 0.951899\n",
      "Precision: 0.914927\n",
      "Recall: 0.951899\n",
      "F1 Score: 0.933047\n",
      "IoU Foreground: 0.874496\n",
      "IoU Background: 0.964158\n",
      "Mean IoU: 0.919327\n",
      "\n",
      "[0.6386279069071205, 0.7643254213630535, 0.7489237904776817, 0.831075799401631, 0.8459055886071709, 0.7921450544677175, 0.7693117470468316, 0.8489570569614452, 0.8528256192134294, 0.8578387507996919, 0.8591140226494551, 0.8744961851984793]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_95_with_prob_0.8744961851984793.pth\n",
      "==========================================================\n",
      "Epoch 2, Itrs 610/10000, Loss=1.156537\n",
      "Epoch 2, Itrs 620/10000, Loss=1.371888\n",
      "Epoch 2, Itrs 630/10000, Loss=1.399238\n",
      "Epoch 2, Itrs 640/10000, Loss=1.244729\n",
      "Epoch 2, Itrs 650/10000, Loss=1.288954\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "124it [00:07, 16.27it/s]\n",
      "Validation Loss: 0.404420\n",
      "\n",
      "Foreground Acc: 0.974926\n",
      "Precision: 0.882283\n",
      "Recall: 0.974926\n",
      "F1 Score: 0.926294\n",
      "IoU Foreground: 0.862707\n",
      "IoU Background: 0.959046\n",
      "Mean IoU: 0.910877\n",
      "\n",
      "[0.6386279069071205, 0.7643254213630535, 0.7489237904776817, 0.831075799401631, 0.8459055886071709, 0.7921450544677175, 0.7693117470468316, 0.8489570569614452, 0.8528256192134294, 0.8578387507996919, 0.8591140226494551, 0.8744961851984793, 0.8627074748604986]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650]\n",
      "Model saved as saved_models/kl_check_temp_2_weighted.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 2, Itrs 660/10000, Loss=1.219869\n",
      "Epoch 2, Itrs 670/10000, Loss=1.274366\n",
      "Epoch 2, Itrs 680/10000, Loss=1.196978\n",
      "Epoch 2, Itrs 690/10000, Loss=1.286288\n",
      "Epoch 2, Itrs 700/10000, Loss=1.279420\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "79it [00:05, 15.64it/s]^C\n",
      "79it [00:05, 15.37it/s]\n",
      "Traceback (most recent call last):\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/binary_leaf_kl_updated.py\", line 647, in <module>\n",
      "    main()\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/binary_leaf_kl_updated.py\", line 589, in main\n",
      "    val_score, ret_samples, current_val_loss = validate(\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/binary_leaf_kl_updated.py\", line 342, in validate\n",
      "    metrics.update(targets, preds)\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/metrics/Binary_metrics.py\", line 24, in update\n",
      "    self.confusion_matrix += self._fast_hist(lt.flatten(), lp.flatten())\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/metrics/Binary_metrics.py\", line 16, in _fast_hist\n",
      "    hist = np.bincount(\n",
      "KeyboardInterrupt\n"
     ]
    }
   ],
   "source": [
    "!python3 binary_leaf_kl_updated.py --model deeplabv3plus_mobilenet --custom_data_path \"crf_sam_annotations_large/\" --batch_size 4 --lr 0.12 --total_itrs 10000 --loss_type kl_dice_combined --val_interval 50  --separable_conv   \n",
    "# /teamspace/studios/this_studio/Deeplab/saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8213798905479425.pth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "67ae809f-aaa5-45a2-86ad-bd14a488068c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device: cpu\n",
      "Dataset: custom, Train set: 187, Val set: 50\n",
      "[!] Retrain\n",
      "Epoch 1, Itrs 10/10000, Loss=0.616262\n",
      "Epoch 1, Itrs 20/10000, Loss=0.558123\n",
      "Epoch 1, Itrs 30/10000, Loss=0.557548\n",
      "Epoch 1, Itrs 40/10000, Loss=0.576444\n",
      "Epoch 2, Itrs 50/10000, Loss=0.589402\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "13it [00:13,  1.05s/it]\n",
      "Validation Loss: 0.628057\n",
      "\n",
      "Foreground Acc: 1.000000\n",
      "Precision: 0.237518\n",
      "Recall: 1.000000\n",
      "F1 Score: 0.383862\n",
      "IoU Foreground: 0.237518\n",
      "IoU Background: 0.000000\n",
      "Mean IoU: 0.118759\n",
      "\n",
      "[0.23751792907714844]\n",
      "[50]\n",
      "Model saved as saved_models/crf_sam_large_freetown_eastafrica_haiti_sam_threshold_970.23751792907714844.pth\n",
      "saved here.\n",
      "==========================================================\n",
      "Epoch 2, Itrs 60/10000, Loss=0.538243\n",
      "^C\n"
     ]
    }
   ],
   "source": [
    "!python3 binary_leaf_main.py --model deeplabv3plus_mobilenet --custom_data_path \"crf_sam_annotations/\" --batch_size 4 --lr 0.12 --total_itrs 10000 --loss_type binary_dice --val_interval 50  --separable_conv   \n",
    "# /teamspace/studios/this_studio/Deeplab/saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8213798905479425.pth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "eebe4a06-089a-45ea-8341-1601e1c05275",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device: cuda\n",
      "Dataset: custom, Train set: 187, Val set: 50\n",
      "[!] Retrain\n",
      "Epoch 1, Itrs 10/10000, Loss=1.217085\n",
      "Epoch 1, Itrs 20/10000, Loss=1.078695\n",
      "Epoch 2, Itrs 30/10000, Loss=1.048480\n",
      "Epoch 2, Itrs 40/10000, Loss=1.048553\n",
      "Epoch 3, Itrs 50/10000, Loss=0.988182\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "13it [00:01, 12.98it/s]\n",
      "Validation Loss: 0.584623\n",
      "\n",
      "Foreground Acc: 0.901423\n",
      "Precision: 0.881515\n",
      "Recall: 0.901423\n",
      "F1 Score: 0.891358\n",
      "IoU Foreground: 0.804009\n",
      "IoU Background: 0.933590\n",
      "Mean IoU: 0.868799\n",
      "\n",
      "[0.8040085468910082]\n",
      "[50]\n",
      "Model saved as saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8040085468910082.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 3, Itrs 60/10000, Loss=0.996446\n",
      "Epoch 4, Itrs 70/10000, Loss=0.971701\n",
      "Epoch 4, Itrs 80/10000, Loss=0.984593\n",
      "Epoch 4, Itrs 90/10000, Loss=0.986145\n",
      "Epoch 5, Itrs 100/10000, Loss=0.947798\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "13it [00:00, 14.47it/s]\n",
      "Validation Loss: 0.555969\n",
      "\n",
      "Foreground Acc: 0.963364\n",
      "Precision: 0.841417\n",
      "Recall: 0.963364\n",
      "F1 Score: 0.898271\n",
      "IoU Foreground: 0.815328\n",
      "IoU Background: 0.932796\n",
      "Mean IoU: 0.874062\n",
      "\n",
      "[0.8040085468910082, 0.8153278230384119]\n",
      "[50, 100]\n",
      "Model saved as saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8153278230384119.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 5, Itrs 110/10000, Loss=0.963851\n",
      "Epoch 6, Itrs 120/10000, Loss=0.947714\n",
      "Epoch 6, Itrs 130/10000, Loss=0.959233\n",
      "Epoch 7, Itrs 140/10000, Loss=0.942581\n",
      "Epoch 7, Itrs 150/10000, Loss=0.959500\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "13it [00:00, 15.72it/s]\n",
      "Validation Loss: 0.466131\n",
      "\n",
      "Foreground Acc: 0.968895\n",
      "Precision: 0.856977\n",
      "Recall: 0.968895\n",
      "F1 Score: 0.909506\n",
      "IoU Foreground: 0.834031\n",
      "IoU Background: 0.940516\n",
      "Mean IoU: 0.887273\n",
      "\n",
      "[0.8040085468910082, 0.8153278230384119, 0.8340307188817794]\n",
      "[50, 100, 150]\n",
      "Model saved as saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8340307188817794.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 7, Itrs 160/10000, Loss=0.946160\n",
      "Epoch 8, Itrs 170/10000, Loss=0.961215\n",
      "Epoch 8, Itrs 180/10000, Loss=0.939007\n",
      "Epoch 9, Itrs 190/10000, Loss=0.915523\n",
      "Epoch 9, Itrs 200/10000, Loss=0.962104\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "13it [00:00, 15.73it/s]\n",
      "Validation Loss: 0.410870\n",
      "\n",
      "Foreground Acc: 0.962323\n",
      "Precision: 0.890021\n",
      "Recall: 0.962323\n",
      "F1 Score: 0.924761\n",
      "IoU Foreground: 0.860051\n",
      "IoU Background: 0.951787\n",
      "Mean IoU: 0.905919\n",
      "\n",
      "[0.8040085468910082, 0.8153278230384119, 0.8340307188817794, 0.8600512258726829]\n",
      "[50, 100, 150, 200]\n",
      "Model saved as saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8600512258726829.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 10, Itrs 210/10000, Loss=0.977649\n",
      "Epoch 10, Itrs 220/10000, Loss=0.933836\n",
      "Epoch 10, Itrs 230/10000, Loss=0.919242\n",
      "Epoch 11, Itrs 240/10000, Loss=0.904043\n",
      "Epoch 11, Itrs 250/10000, Loss=0.949915\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "13it [00:00, 16.01it/s]\n",
      "Validation Loss: 0.393625\n",
      "\n",
      "Foreground Acc: 0.963671\n",
      "Precision: 0.897391\n",
      "Recall: 0.963671\n",
      "F1 Score: 0.929351\n",
      "IoU Foreground: 0.868026\n",
      "IoU Background: 0.954870\n",
      "Mean IoU: 0.911448\n",
      "\n",
      "[0.8040085468910082, 0.8153278230384119, 0.8340307188817794, 0.8600512258726829, 0.8680256036127215]\n",
      "[50, 100, 150, 200, 250]\n",
      "Model saved as saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8680256036127215.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 12, Itrs 260/10000, Loss=0.922074\n",
      "Epoch 12, Itrs 270/10000, Loss=0.892848\n",
      "Epoch 13, Itrs 280/10000, Loss=0.920650\n",
      "Epoch 13, Itrs 290/10000, Loss=0.952084\n",
      "Epoch 14, Itrs 300/10000, Loss=0.924332\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "13it [00:00, 15.86it/s]\n",
      "Validation Loss: 0.347076\n",
      "\n",
      "Foreground Acc: 0.969130\n",
      "Precision: 0.894251\n",
      "Recall: 0.969130\n",
      "F1 Score: 0.930186\n",
      "IoU Foreground: 0.869484\n",
      "IoU Background: 0.955116\n",
      "Mean IoU: 0.912300\n",
      "\n",
      "[0.8040085468910082, 0.8153278230384119, 0.8340307188817794, 0.8600512258726829, 0.8680256036127215, 0.8694836481719341]\n",
      "[50, 100, 150, 200, 250, 300]\n",
      "Model saved as saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8694836481719341.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 14, Itrs 310/10000, Loss=0.940602\n",
      "Epoch 14, Itrs 320/10000, Loss=0.930540\n",
      "Epoch 15, Itrs 330/10000, Loss=0.917509\n",
      "Epoch 15, Itrs 340/10000, Loss=0.922430\n",
      "Epoch 16, Itrs 350/10000, Loss=0.948368\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "13it [00:00, 16.11it/s]\n",
      "Validation Loss: 0.493670\n",
      "\n",
      "Foreground Acc: 0.951571\n",
      "Precision: 0.916238\n",
      "Recall: 0.951571\n",
      "F1 Score: 0.933571\n",
      "IoU Foreground: 0.875417\n",
      "IoU Background: 0.958443\n",
      "Mean IoU: 0.916930\n",
      "\n",
      "[0.8040085468910082, 0.8153278230384119, 0.8340307188817794, 0.8600512258726829, 0.8680256036127215, 0.8694836481719341, 0.8754172925254217]\n",
      "[50, 100, 150, 200, 250, 300, 350]\n",
      "Model saved as saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8754172925254217.pth\n",
      "saved here\n",
      "==========================================================\n",
      "Epoch 16, Itrs 360/10000, Loss=0.901046\n",
      "Epoch 17, Itrs 370/10000, Loss=0.930715\n",
      "^C\n",
      "Traceback (most recent call last):\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/binary_leaf_main.py\", line 596, in <module>\n",
      "    main()\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/binary_leaf_main.py\", line 523, in main\n",
      "    np_loss = loss.detach().cpu().numpy()\n",
      "KeyboardInterrupt\n"
     ]
    }
   ],
   "source": [
    "!python3 binary_leaf_main.py --model deeplabv3plus_mobilenet --custom_data_path \"crf_sam_annotations/\" --batch_size 8 --lr 0.12 --total_itrs 10000 --loss_type binary_dice --val_interval 50  --separable_conv   \n",
    "# /teamspace/studios/this_studio/Deeplab/saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8213798905479425.pth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b7253a65-e553-4a0c-aa36-a92497d3b5d8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device: cpu\n",
      "Dataset: custom, Train set: 48, Val set: 183\n",
      "Model restored from /teamspace/studios/this_studio/Deeplab/saved_models/india_sam_dino_250_samples_0.8940880817269634.pth\n",
      "46it [01:05,  1.42s/it]\n",
      "\n",
      "Foreground Acc: 0.920715\n",
      "Precision: 0.743458\n",
      "Recall: 0.920715\n",
      "F1 Score: 0.822647\n",
      "IoU Foreground: 0.698726\n",
      "IoU Background: 0.929017\n",
      "Mean IoU: 0.813871\n",
      "\n"
     ]
    }
   ],
   "source": [
    "!python3 binary_leaf_main.py --model deeplabv3plus_mobilenet --custom_data_path \"pilot_annotations_eastafrica_haiti_freetown/\" --batch_size 16 --lr 0.01 --total_itrs 10000 --loss_type binary_dice --val_interval 50  --separable_conv --ckpt $downstreamed_small\n",
    "#/teamspace/studios/this_studio/Deeplab/saved_models/base_sam_dino_india_tuned_freetown_eastafrica_haiti_0.8213798905479425.pth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "01c66858-1456-41cd-895d-346f54c25c62",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device: cuda\n",
      "Dataset: custom, Train set: 3053, Val set: 814\n",
      "[!] Retrain\n",
      "Epoch 1, Itrs 10/10000, Loss=1.327401\n",
      "Epoch 1, Itrs 20/10000, Loss=1.142678\n",
      "Epoch 1, Itrs 30/10000, Loss=1.074712\n",
      "Epoch 1, Itrs 40/10000, Loss=1.067894\n",
      "Epoch 1, Itrs 50/10000, Loss=1.052424\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.18it/s]\n",
      "Validation Loss: 0.785874\n",
      "\n",
      "Foreground Acc: 0.529455\n",
      "Precision: 0.863823\n",
      "Recall: 0.529455\n",
      "F1 Score: 0.656517\n",
      "IoU Foreground: 0.488668\n",
      "IoU Background: 0.884323\n",
      "Mean IoU: 0.686495\n",
      "\n",
      "[0.4886676694697532]\n",
      "[50]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.4886676694697532.pth\n",
      "==========================================================\n",
      "Epoch 1, Itrs 60/10000, Loss=1.042793\n",
      "Epoch 1, Itrs 70/10000, Loss=1.031934\n",
      "Epoch 1, Itrs 80/10000, Loss=1.014307\n",
      "Epoch 1, Itrs 90/10000, Loss=1.011342\n",
      "Epoch 1, Itrs 100/10000, Loss=1.011995\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.24it/s]\n",
      "Validation Loss: 0.639642\n",
      "\n",
      "Foreground Acc: 0.778738\n",
      "Precision: 0.845785\n",
      "Recall: 0.778738\n",
      "F1 Score: 0.810878\n",
      "IoU Foreground: 0.681913\n",
      "IoU Background: 0.919988\n",
      "Mean IoU: 0.800951\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402]\n",
      "[50, 100]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.6819128100575402.pth\n",
      "==========================================================\n",
      "Epoch 1, Itrs 110/10000, Loss=0.998304\n",
      "Epoch 1, Itrs 120/10000, Loss=0.994156\n",
      "Epoch 1, Itrs 130/10000, Loss=0.987849\n",
      "Epoch 1, Itrs 140/10000, Loss=0.983608\n",
      "Epoch 1, Itrs 150/10000, Loss=0.987636\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.17it/s]\n",
      "Validation Loss: 0.610341\n",
      "\n",
      "Foreground Acc: 0.840431\n",
      "Precision: 0.849721\n",
      "Recall: 0.840431\n",
      "F1 Score: 0.845051\n",
      "IoU Foreground: 0.731678\n",
      "IoU Background: 0.931178\n",
      "Mean IoU: 0.831428\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185]\n",
      "[50, 100, 150]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.7316776324091185.pth\n",
      "==========================================================\n",
      "Epoch 1, Itrs 160/10000, Loss=0.997143\n",
      "Epoch 1, Itrs 170/10000, Loss=0.994399\n",
      "Epoch 1, Itrs 180/10000, Loss=1.001625\n",
      "Epoch 1, Itrs 190/10000, Loss=0.969644\n",
      "Epoch 2, Itrs 200/10000, Loss=0.988063\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.05it/s]\n",
      "Validation Loss: 0.590060\n",
      "\n",
      "Foreground Acc: 0.882715\n",
      "Precision: 0.837912\n",
      "Recall: 0.882715\n",
      "F1 Score: 0.859730\n",
      "IoU Foreground: 0.753970\n",
      "IoU Background: 0.935068\n",
      "Mean IoU: 0.844519\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451]\n",
      "[50, 100, 150, 200]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.7539703610134451.pth\n",
      "==========================================================\n",
      "Epoch 2, Itrs 210/10000, Loss=0.980325\n",
      "Epoch 2, Itrs 220/10000, Loss=0.979616\n",
      "Epoch 2, Itrs 230/10000, Loss=0.973685\n",
      "Epoch 2, Itrs 240/10000, Loss=0.972328\n",
      "Epoch 2, Itrs 250/10000, Loss=0.967510\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.10it/s]\n",
      "Validation Loss: 0.573927\n",
      "\n",
      "Foreground Acc: 0.902157\n",
      "Precision: 0.833825\n",
      "Recall: 0.902157\n",
      "F1 Score: 0.866646\n",
      "IoU Foreground: 0.764673\n",
      "IoU Background: 0.937138\n",
      "Mean IoU: 0.850905\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697]\n",
      "[50, 100, 150, 200, 250]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.7646733691379697.pth\n",
      "==========================================================\n",
      "Epoch 2, Itrs 260/10000, Loss=0.967980\n",
      "Epoch 2, Itrs 270/10000, Loss=0.979676\n",
      "Epoch 2, Itrs 280/10000, Loss=0.970957\n",
      "Epoch 2, Itrs 290/10000, Loss=0.957425\n",
      "Epoch 2, Itrs 300/10000, Loss=0.965152\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.32it/s]\n",
      "Validation Loss: 0.561150\n",
      "\n",
      "Foreground Acc: 0.917462\n",
      "Precision: 0.829244\n",
      "Recall: 0.917462\n",
      "F1 Score: 0.871126\n",
      "IoU Foreground: 0.771676\n",
      "IoU Background: 0.938323\n",
      "Mean IoU: 0.855000\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642]\n",
      "[50, 100, 150, 200, 250, 300]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.7716763754696642.pth\n",
      "==========================================================\n",
      "Epoch 2, Itrs 310/10000, Loss=0.974408\n",
      "Epoch 2, Itrs 320/10000, Loss=0.973860\n",
      "Epoch 2, Itrs 330/10000, Loss=0.995087\n",
      "Epoch 2, Itrs 340/10000, Loss=0.951047\n",
      "Epoch 2, Itrs 350/10000, Loss=0.965069\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.23it/s]\n",
      "Validation Loss: 0.552150\n",
      "\n",
      "Foreground Acc: 0.919061\n",
      "Precision: 0.838619\n",
      "Recall: 0.919061\n",
      "F1 Score: 0.876999\n",
      "IoU Foreground: 0.780943\n",
      "IoU Background: 0.941405\n",
      "Mean IoU: 0.861174\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218]\n",
      "[50, 100, 150, 200, 250, 300, 350]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.780942755510218.pth\n",
      "==========================================================\n",
      "Epoch 2, Itrs 360/10000, Loss=0.942939\n",
      "Epoch 2, Itrs 370/10000, Loss=0.949125\n",
      "Epoch 2, Itrs 380/10000, Loss=0.976573\n",
      "Epoch 3, Itrs 390/10000, Loss=0.949162\n",
      "Epoch 3, Itrs 400/10000, Loss=0.963617\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.29it/s]\n",
      "Validation Loss: 0.551177\n",
      "\n",
      "Foreground Acc: 0.917127\n",
      "Precision: 0.849755\n",
      "Recall: 0.917127\n",
      "F1 Score: 0.882156\n",
      "IoU Foreground: 0.789159\n",
      "IoU Background: 0.944332\n",
      "Mean IoU: 0.866745\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.7891591265755157.pth\n",
      "==========================================================\n",
      "Epoch 3, Itrs 410/10000, Loss=0.967226\n",
      "Epoch 3, Itrs 420/10000, Loss=0.964278\n",
      "Epoch 3, Itrs 430/10000, Loss=0.950411\n",
      "Epoch 3, Itrs 440/10000, Loss=0.945423\n",
      "Epoch 3, Itrs 450/10000, Loss=0.965909\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.09it/s]\n",
      "Validation Loss: 0.536343\n",
      "\n",
      "Foreground Acc: 0.920936\n",
      "Precision: 0.852973\n",
      "Recall: 0.920936\n",
      "F1 Score: 0.885653\n",
      "IoU Foreground: 0.794772\n",
      "IoU Background: 0.945926\n",
      "Mean IoU: 0.870349\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.7947723249076609.pth\n",
      "==========================================================\n",
      "Epoch 3, Itrs 460/10000, Loss=0.956756\n",
      "Epoch 3, Itrs 470/10000, Loss=0.961028\n",
      "Epoch 3, Itrs 480/10000, Loss=0.971151\n",
      "Epoch 3, Itrs 490/10000, Loss=0.951505\n",
      "Epoch 3, Itrs 500/10000, Loss=0.940505\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.03it/s]\n",
      "Validation Loss: 0.533632\n",
      "\n",
      "Foreground Acc: 0.926853\n",
      "Precision: 0.851290\n",
      "Recall: 0.926853\n",
      "F1 Score: 0.887466\n",
      "IoU Foreground: 0.797698\n",
      "IoU Background: 0.946479\n",
      "Mean IoU: 0.872089\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.7976980095166158.pth\n",
      "==========================================================\n",
      "Epoch 3, Itrs 510/10000, Loss=0.942220\n",
      "Epoch 3, Itrs 520/10000, Loss=0.946827\n",
      "Epoch 3, Itrs 530/10000, Loss=0.953365\n",
      "Epoch 3, Itrs 540/10000, Loss=0.944968\n",
      "Epoch 3, Itrs 550/10000, Loss=0.950847\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.20it/s]\n",
      "Validation Loss: 0.532581\n",
      "\n",
      "Foreground Acc: 0.928332\n",
      "Precision: 0.855893\n",
      "Recall: 0.928332\n",
      "F1 Score: 0.890642\n",
      "IoU Foreground: 0.802845\n",
      "IoU Background: 0.948075\n",
      "Mean IoU: 0.875460\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.80284455982853.pth\n",
      "==========================================================\n",
      "Epoch 3, Itrs 560/10000, Loss=0.950642\n",
      "Epoch 3, Itrs 570/10000, Loss=0.955682\n",
      "Epoch 4, Itrs 580/10000, Loss=0.956354\n",
      "Epoch 4, Itrs 590/10000, Loss=0.947548\n",
      "Epoch 4, Itrs 600/10000, Loss=0.944109\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.20it/s]\n",
      "Validation Loss: 0.529665\n",
      "\n",
      "Foreground Acc: 0.940016\n",
      "Precision: 0.842584\n",
      "Recall: 0.940016\n",
      "F1 Score: 0.888637\n",
      "IoU Foreground: 0.799593\n",
      "IoU Background: 0.946194\n",
      "Mean IoU: 0.872893\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600]\n",
      "==========================================================\n",
      "Epoch 4, Itrs 610/10000, Loss=0.932166\n",
      "Epoch 4, Itrs 620/10000, Loss=0.946571\n",
      "Epoch 4, Itrs 630/10000, Loss=0.952019\n",
      "Epoch 4, Itrs 640/10000, Loss=0.934304\n",
      "Epoch 4, Itrs 650/10000, Loss=0.932865\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.21it/s]\n",
      "Validation Loss: 0.528190\n",
      "\n",
      "Foreground Acc: 0.934959\n",
      "Precision: 0.857271\n",
      "Recall: 0.934959\n",
      "F1 Score: 0.894431\n",
      "IoU Foreground: 0.809023\n",
      "IoU Background: 0.949654\n",
      "Mean IoU: 0.879339\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.8090232975237008.pth\n",
      "==========================================================\n",
      "Epoch 4, Itrs 660/10000, Loss=0.966130\n",
      "Epoch 4, Itrs 670/10000, Loss=0.935611\n",
      "Epoch 4, Itrs 680/10000, Loss=0.930244\n",
      "Epoch 4, Itrs 690/10000, Loss=0.932908\n",
      "Epoch 4, Itrs 700/10000, Loss=0.953318\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.38it/s]\n",
      "Validation Loss: 0.513412\n",
      "\n",
      "Foreground Acc: 0.927282\n",
      "Precision: 0.870538\n",
      "Recall: 0.927282\n",
      "F1 Score: 0.898014\n",
      "IoU Foreground: 0.814906\n",
      "IoU Background: 0.952039\n",
      "Mean IoU: 0.883472\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.8149058771937127.pth\n",
      "==========================================================\n",
      "Epoch 4, Itrs 710/10000, Loss=0.934967\n",
      "Epoch 4, Itrs 720/10000, Loss=0.939226\n",
      "Epoch 4, Itrs 730/10000, Loss=0.939435\n",
      "Epoch 4, Itrs 740/10000, Loss=0.936865\n",
      "Epoch 4, Itrs 750/10000, Loss=0.944344\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.13it/s]\n",
      "Validation Loss: 0.516123\n",
      "\n",
      "Foreground Acc: 0.939428\n",
      "Precision: 0.857610\n",
      "Recall: 0.939428\n",
      "F1 Score: 0.896656\n",
      "IoU Foreground: 0.812672\n",
      "IoU Background: 0.950552\n",
      "Mean IoU: 0.881612\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750]\n",
      "==========================================================\n",
      "Epoch 4, Itrs 760/10000, Loss=0.952620\n",
      "Epoch 5, Itrs 770/10000, Loss=0.931900\n",
      "Epoch 5, Itrs 780/10000, Loss=0.941550\n",
      "Epoch 5, Itrs 790/10000, Loss=0.931792\n",
      "Epoch 5, Itrs 800/10000, Loss=0.939280\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:12, 16.87it/s]\n",
      "Validation Loss: 0.502783\n",
      "\n",
      "Foreground Acc: 0.933923\n",
      "Precision: 0.868797\n",
      "Recall: 0.933923\n",
      "F1 Score: 0.900184\n",
      "IoU Foreground: 0.818485\n",
      "IoU Background: 0.952765\n",
      "Mean IoU: 0.885625\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.8184854797204618.pth\n",
      "==========================================================\n",
      "Epoch 5, Itrs 810/10000, Loss=0.935501\n",
      "Epoch 5, Itrs 820/10000, Loss=0.929339\n",
      "Epoch 5, Itrs 830/10000, Loss=0.946889\n",
      "Epoch 5, Itrs 840/10000, Loss=0.941112\n",
      "Epoch 5, Itrs 850/10000, Loss=0.936314\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.22it/s]\n",
      "Validation Loss: 0.508537\n",
      "\n",
      "Foreground Acc: 0.941636\n",
      "Precision: 0.861781\n",
      "Recall: 0.941636\n",
      "F1 Score: 0.899940\n",
      "IoU Foreground: 0.818083\n",
      "IoU Background: 0.952162\n",
      "Mean IoU: 0.885123\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850]\n",
      "==========================================================\n",
      "Epoch 5, Itrs 860/10000, Loss=0.952079\n",
      "Epoch 5, Itrs 870/10000, Loss=0.923357\n",
      "Epoch 5, Itrs 880/10000, Loss=0.935570\n",
      "Epoch 5, Itrs 890/10000, Loss=0.941258\n",
      "Epoch 5, Itrs 900/10000, Loss=0.933962\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.17it/s]\n",
      "Validation Loss: 0.508583\n",
      "\n",
      "Foreground Acc: 0.945600\n",
      "Precision: 0.858887\n",
      "Recall: 0.945600\n",
      "F1 Score: 0.900160\n",
      "IoU Foreground: 0.818446\n",
      "IoU Background: 0.952034\n",
      "Mean IoU: 0.885240\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900]\n",
      "==========================================================\n",
      "Epoch 5, Itrs 910/10000, Loss=0.943520\n",
      "Epoch 5, Itrs 920/10000, Loss=0.944898\n",
      "Epoch 5, Itrs 930/10000, Loss=0.930383\n",
      "Epoch 5, Itrs 940/10000, Loss=0.947989\n",
      "Epoch 5, Itrs 950/10000, Loss=0.934782\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.22it/s]\n",
      "Validation Loss: 0.524006\n",
      "\n",
      "Foreground Acc: 0.951769\n",
      "Precision: 0.849997\n",
      "Recall: 0.951769\n",
      "F1 Score: 0.898009\n",
      "IoU Foreground: 0.814897\n",
      "IoU Background: 0.950493\n",
      "Mean IoU: 0.882695\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950]\n",
      "==========================================================\n",
      "Epoch 6, Itrs 960/10000, Loss=0.921053\n",
      "Epoch 6, Itrs 970/10000, Loss=0.937063\n",
      "Epoch 6, Itrs 980/10000, Loss=0.943824\n",
      "Epoch 6, Itrs 990/10000, Loss=0.916251\n",
      "Epoch 6, Itrs 1000/10000, Loss=0.933443\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.15it/s]\n",
      "Validation Loss: 0.500019\n",
      "\n",
      "Foreground Acc: 0.941077\n",
      "Precision: 0.871613\n",
      "Recall: 0.941077\n",
      "F1 Score: 0.905014\n",
      "IoU Foreground: 0.826507\n",
      "IoU Background: 0.954875\n",
      "Mean IoU: 0.890691\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.8265067495056632.pth\n",
      "==========================================================\n",
      "Epoch 6, Itrs 1010/10000, Loss=0.935302\n",
      "Epoch 6, Itrs 1020/10000, Loss=0.924963\n",
      "Epoch 6, Itrs 1030/10000, Loss=0.944003\n",
      "Epoch 6, Itrs 1040/10000, Loss=0.942149\n",
      "Epoch 6, Itrs 1050/10000, Loss=0.934149\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.06it/s]\n",
      "Validation Loss: 0.496212\n",
      "\n",
      "Foreground Acc: 0.951569\n",
      "Precision: 0.855449\n",
      "Recall: 0.951569\n",
      "F1 Score: 0.900952\n",
      "IoU Foreground: 0.819758\n",
      "IoU Background: 0.952091\n",
      "Mean IoU: 0.885925\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050]\n",
      "==========================================================\n",
      "Epoch 6, Itrs 1060/10000, Loss=0.932647\n",
      "Epoch 6, Itrs 1070/10000, Loss=0.937932\n",
      "Epoch 6, Itrs 1080/10000, Loss=0.939175\n",
      "Epoch 6, Itrs 1090/10000, Loss=0.927619\n",
      "Epoch 6, Itrs 1100/10000, Loss=0.926780\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.21it/s]\n",
      "Validation Loss: 0.501504\n",
      "\n",
      "Foreground Acc: 0.945780\n",
      "Precision: 0.868979\n",
      "Recall: 0.945780\n",
      "F1 Score: 0.905754\n",
      "IoU Foreground: 0.827743\n",
      "IoU Background: 0.954991\n",
      "Mean IoU: 0.891367\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.8277434022634125.pth\n",
      "==========================================================\n",
      "Epoch 6, Itrs 1110/10000, Loss=0.928373\n",
      "Epoch 6, Itrs 1120/10000, Loss=0.924923\n",
      "Epoch 6, Itrs 1130/10000, Loss=0.938881\n",
      "Epoch 6, Itrs 1140/10000, Loss=0.930399\n",
      "Epoch 7, Itrs 1150/10000, Loss=0.920155\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.28it/s]\n",
      "Validation Loss: 0.504543\n",
      "\n",
      "Foreground Acc: 0.956695\n",
      "Precision: 0.849626\n",
      "Recall: 0.956695\n",
      "F1 Score: 0.899987\n",
      "IoU Foreground: 0.818161\n",
      "IoU Background: 0.951255\n",
      "Mean IoU: 0.884708\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150]\n",
      "==========================================================\n",
      "Epoch 7, Itrs 1160/10000, Loss=0.936380\n",
      "Epoch 7, Itrs 1170/10000, Loss=0.958067\n",
      "Epoch 7, Itrs 1180/10000, Loss=0.910540\n",
      "Epoch 7, Itrs 1190/10000, Loss=0.912680\n",
      "Epoch 7, Itrs 1200/10000, Loss=0.932913\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.33it/s]\n",
      "Validation Loss: 0.500886\n",
      "\n",
      "Foreground Acc: 0.955804\n",
      "Precision: 0.854987\n",
      "Recall: 0.955804\n",
      "F1 Score: 0.902589\n",
      "IoU Foreground: 0.822471\n",
      "IoU Background: 0.952713\n",
      "Mean IoU: 0.887592\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200]\n",
      "==========================================================\n",
      "Epoch 7, Itrs 1210/10000, Loss=0.928441\n",
      "Epoch 7, Itrs 1220/10000, Loss=0.938010\n",
      "Epoch 7, Itrs 1230/10000, Loss=0.925875\n",
      "Epoch 7, Itrs 1240/10000, Loss=0.939202\n",
      "Epoch 7, Itrs 1250/10000, Loss=0.936043\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.20it/s]\n",
      "Validation Loss: 0.502165\n",
      "\n",
      "Foreground Acc: 0.955221\n",
      "Precision: 0.857486\n",
      "Recall: 0.955221\n",
      "F1 Score: 0.903719\n",
      "IoU Foreground: 0.824349\n",
      "IoU Background: 0.953355\n",
      "Mean IoU: 0.888852\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250]\n",
      "==========================================================\n",
      "Epoch 7, Itrs 1260/10000, Loss=0.937954\n",
      "Epoch 7, Itrs 1270/10000, Loss=0.905312\n",
      "Epoch 7, Itrs 1280/10000, Loss=0.942084\n",
      "Epoch 7, Itrs 1290/10000, Loss=0.938045\n",
      "Epoch 7, Itrs 1300/10000, Loss=0.913250\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.02it/s]\n",
      "Validation Loss: 0.495878\n",
      "\n",
      "Foreground Acc: 0.954743\n",
      "Precision: 0.860027\n",
      "Recall: 0.954743\n",
      "F1 Score: 0.904913\n",
      "IoU Foreground: 0.826340\n",
      "IoU Background: 0.954023\n",
      "Mean IoU: 0.890181\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300]\n",
      "==========================================================\n",
      "Epoch 7, Itrs 1310/10000, Loss=0.940445\n",
      "Epoch 7, Itrs 1320/10000, Loss=0.937134\n",
      "Epoch 7, Itrs 1330/10000, Loss=0.916464\n",
      "Epoch 8, Itrs 1340/10000, Loss=0.916476\n",
      "Epoch 8, Itrs 1350/10000, Loss=0.932945\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.18it/s]\n",
      "Validation Loss: 0.489941\n",
      "\n",
      "Foreground Acc: 0.952336\n",
      "Precision: 0.868187\n",
      "Recall: 0.952336\n",
      "F1 Score: 0.908317\n",
      "IoU Foreground: 0.832033\n",
      "IoU Background: 0.955970\n",
      "Mean IoU: 0.894001\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.8320330831656192.pth\n",
      "==========================================================\n",
      "Epoch 8, Itrs 1360/10000, Loss=0.916665\n",
      "Epoch 8, Itrs 1370/10000, Loss=0.945384\n",
      "Epoch 8, Itrs 1380/10000, Loss=0.901965\n",
      "Epoch 8, Itrs 1390/10000, Loss=0.920679\n",
      "Epoch 8, Itrs 1400/10000, Loss=0.928654\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.21it/s]\n",
      "Validation Loss: 0.488884\n",
      "\n",
      "Foreground Acc: 0.952443\n",
      "Precision: 0.869503\n",
      "Recall: 0.952443\n",
      "F1 Score: 0.909085\n",
      "IoU Foreground: 0.833323\n",
      "IoU Background: 0.956370\n",
      "Mean IoU: 0.894847\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.8333233742814653.pth\n",
      "==========================================================\n",
      "Epoch 8, Itrs 1410/10000, Loss=0.927316\n",
      "Epoch 8, Itrs 1420/10000, Loss=0.924522\n",
      "Epoch 8, Itrs 1430/10000, Loss=0.939777\n",
      "Epoch 8, Itrs 1440/10000, Loss=0.940097\n",
      "Epoch 8, Itrs 1450/10000, Loss=0.910281\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.21it/s]\n",
      "Validation Loss: 0.491126\n",
      "\n",
      "Foreground Acc: 0.957366\n",
      "Precision: 0.860151\n",
      "Recall: 0.957366\n",
      "F1 Score: 0.906159\n",
      "IoU Foreground: 0.828418\n",
      "IoU Background: 0.954535\n",
      "Mean IoU: 0.891477\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450]\n",
      "==========================================================\n",
      "Epoch 8, Itrs 1460/10000, Loss=0.945357\n",
      "Epoch 8, Itrs 1470/10000, Loss=0.908695\n",
      "Epoch 8, Itrs 1480/10000, Loss=0.933841\n",
      "Epoch 8, Itrs 1490/10000, Loss=0.915697\n",
      "Epoch 8, Itrs 1500/10000, Loss=0.926782\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.22it/s]\n",
      "Validation Loss: 0.483421\n",
      "\n",
      "Foreground Acc: 0.951322\n",
      "Precision: 0.874195\n",
      "Recall: 0.951322\n",
      "F1 Score: 0.911129\n",
      "IoU Foreground: 0.836765\n",
      "IoU Background: 0.957507\n",
      "Mean IoU: 0.897136\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.8367651212779491.pth\n",
      "==========================================================\n",
      "Epoch 8, Itrs 1510/10000, Loss=0.927467\n",
      "Epoch 8, Itrs 1520/10000, Loss=0.917349\n",
      "Epoch 9, Itrs 1530/10000, Loss=0.908592\n",
      "Epoch 9, Itrs 1540/10000, Loss=0.941675\n",
      "Epoch 9, Itrs 1550/10000, Loss=0.922193\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:12, 16.95it/s]\n",
      "Validation Loss: 0.485986\n",
      "\n",
      "Foreground Acc: 0.956436\n",
      "Precision: 0.865003\n",
      "Recall: 0.956436\n",
      "F1 Score: 0.908424\n",
      "IoU Foreground: 0.832214\n",
      "IoU Background: 0.955796\n",
      "Mean IoU: 0.894005\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550]\n",
      "==========================================================\n",
      "Epoch 9, Itrs 1560/10000, Loss=0.918170\n",
      "Epoch 9, Itrs 1570/10000, Loss=0.920977\n",
      "Epoch 9, Itrs 1580/10000, Loss=0.930802\n",
      "Epoch 9, Itrs 1590/10000, Loss=0.920885\n",
      "Epoch 9, Itrs 1600/10000, Loss=0.923051\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:12, 16.78it/s]\n",
      "Validation Loss: 0.490311\n",
      "\n",
      "Foreground Acc: 0.955901\n",
      "Precision: 0.869012\n",
      "Recall: 0.955901\n",
      "F1 Score: 0.910388\n",
      "IoU Foreground: 0.835516\n",
      "IoU Background: 0.956867\n",
      "Mean IoU: 0.896191\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600]\n",
      "==========================================================\n",
      "Epoch 9, Itrs 1610/10000, Loss=0.918038\n",
      "Epoch 9, Itrs 1620/10000, Loss=0.923723\n",
      "Epoch 9, Itrs 1630/10000, Loss=0.935374\n",
      "Epoch 9, Itrs 1640/10000, Loss=0.914865\n",
      "Epoch 9, Itrs 1650/10000, Loss=0.933209\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.22it/s]\n",
      "Validation Loss: 0.481440\n",
      "\n",
      "Foreground Acc: 0.951942\n",
      "Precision: 0.877360\n",
      "Recall: 0.951942\n",
      "F1 Score: 0.913131\n",
      "IoU Foreground: 0.840148\n",
      "IoU Background: 0.958523\n",
      "Mean IoU: 0.899335\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.8401478523183005.pth\n",
      "==========================================================\n",
      "Epoch 9, Itrs 1660/10000, Loss=0.921907\n",
      "Epoch 9, Itrs 1670/10000, Loss=0.933911\n",
      "Epoch 9, Itrs 1680/10000, Loss=0.918896\n",
      "Epoch 9, Itrs 1690/10000, Loss=0.914870\n",
      "Epoch 9, Itrs 1700/10000, Loss=0.937473\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.05it/s]\n",
      "Validation Loss: 0.483658\n",
      "\n",
      "Foreground Acc: 0.952232\n",
      "Precision: 0.878468\n",
      "Recall: 0.952232\n",
      "F1 Score: 0.913864\n",
      "IoU Foreground: 0.841390\n",
      "IoU Background: 0.958890\n",
      "Mean IoU: 0.900140\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.8413897342847224.pth\n",
      "==========================================================\n",
      "Epoch 9, Itrs 1710/10000, Loss=0.921523\n",
      "Epoch 10, Itrs 1720/10000, Loss=0.940322\n",
      "Epoch 10, Itrs 1730/10000, Loss=0.924496\n",
      "Epoch 10, Itrs 1740/10000, Loss=0.924432\n",
      "Epoch 10, Itrs 1750/10000, Loss=0.910524\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.07it/s]\n",
      "Validation Loss: 0.477842\n",
      "\n",
      "Foreground Acc: 0.956142\n",
      "Precision: 0.871168\n",
      "Recall: 0.956142\n",
      "F1 Score: 0.911679\n",
      "IoU Foreground: 0.837693\n",
      "IoU Background: 0.957535\n",
      "Mean IoU: 0.897614\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224, 0.8376933561199469]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700, 1750]\n",
      "==========================================================\n",
      "Epoch 10, Itrs 1760/10000, Loss=0.924834\n",
      "Epoch 10, Itrs 1770/10000, Loss=0.921462\n",
      "Epoch 10, Itrs 1780/10000, Loss=0.906448\n",
      "Epoch 10, Itrs 1790/10000, Loss=0.918895\n",
      "Epoch 10, Itrs 1800/10000, Loss=0.895157\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.25it/s]\n",
      "Validation Loss: 0.481665\n",
      "\n",
      "Foreground Acc: 0.958002\n",
      "Precision: 0.867749\n",
      "Recall: 0.958002\n",
      "F1 Score: 0.910645\n",
      "IoU Foreground: 0.835949\n",
      "IoU Background: 0.956887\n",
      "Mean IoU: 0.896418\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224, 0.8376933561199469, 0.8359486673807737]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700, 1750, 1800]\n",
      "==========================================================\n",
      "Epoch 10, Itrs 1810/10000, Loss=0.902858\n",
      "Epoch 10, Itrs 1820/10000, Loss=0.941139\n",
      "Epoch 10, Itrs 1830/10000, Loss=0.924136\n",
      "Epoch 10, Itrs 1840/10000, Loss=0.916695\n",
      "Epoch 10, Itrs 1850/10000, Loss=0.918335\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.19it/s]\n",
      "Validation Loss: 0.471127\n",
      "\n",
      "Foreground Acc: 0.955192\n",
      "Precision: 0.876277\n",
      "Recall: 0.955192\n",
      "F1 Score: 0.914035\n",
      "IoU Foreground: 0.841679\n",
      "IoU Background: 0.958824\n",
      "Mean IoU: 0.900252\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224, 0.8376933561199469, 0.8359486673807737, 0.8416792515202142]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700, 1750, 1800, 1850]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.8416792515202142.pth\n",
      "==========================================================\n",
      "Epoch 10, Itrs 1860/10000, Loss=0.938431\n",
      "Epoch 10, Itrs 1870/10000, Loss=0.940251\n",
      "Epoch 10, Itrs 1880/10000, Loss=0.919944\n",
      "Epoch 10, Itrs 1890/10000, Loss=0.895698\n",
      "Epoch 10, Itrs 1900/10000, Loss=0.927250\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.08it/s]\n",
      "Validation Loss: 0.475727\n",
      "\n",
      "Foreground Acc: 0.944364\n",
      "Precision: 0.894276\n",
      "Recall: 0.944364\n",
      "F1 Score: 0.918638\n",
      "IoU Foreground: 0.849519\n",
      "IoU Background: 0.961759\n",
      "Mean IoU: 0.905639\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224, 0.8376933561199469, 0.8359486673807737, 0.8416792515202142, 0.8495187035894669]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700, 1750, 1800, 1850, 1900]\n",
      "Model saved as saved_models/india_sam_dino_250_samples_0.8495187035894669.pth\n",
      "==========================================================\n",
      "Epoch 11, Itrs 1910/10000, Loss=0.925731\n",
      "Epoch 11, Itrs 1920/10000, Loss=0.940448\n",
      "Epoch 11, Itrs 1930/10000, Loss=0.917090\n",
      "Epoch 11, Itrs 1940/10000, Loss=0.936614\n",
      "Epoch 11, Itrs 1950/10000, Loss=0.918959\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.11it/s]\n",
      "Validation Loss: 0.482967\n",
      "\n",
      "Foreground Acc: 0.959355\n",
      "Precision: 0.870373\n",
      "Recall: 0.959355\n",
      "F1 Score: 0.912700\n",
      "IoU Foreground: 0.839420\n",
      "IoU Background: 0.957901\n",
      "Mean IoU: 0.898660\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224, 0.8376933561199469, 0.8359486673807737, 0.8416792515202142, 0.8495187035894669, 0.8394195386637451]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700, 1750, 1800, 1850, 1900, 1950]\n",
      "==========================================================\n",
      "Epoch 11, Itrs 1960/10000, Loss=0.929936\n",
      "Epoch 11, Itrs 1970/10000, Loss=0.899790\n",
      "Epoch 11, Itrs 1980/10000, Loss=0.925989\n",
      "Epoch 11, Itrs 1990/10000, Loss=0.918928\n",
      "Epoch 11, Itrs 2000/10000, Loss=0.938837\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.07it/s]\n",
      "Validation Loss: 0.483436\n",
      "\n",
      "Foreground Acc: 0.960874\n",
      "Precision: 0.868391\n",
      "Recall: 0.960874\n",
      "F1 Score: 0.912294\n",
      "IoU Foreground: 0.838733\n",
      "IoU Background: 0.957605\n",
      "Mean IoU: 0.898169\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224, 0.8376933561199469, 0.8359486673807737, 0.8416792515202142, 0.8495187035894669, 0.8394195386637451, 0.8387325560736214]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700, 1750, 1800, 1850, 1900, 1950, 2000]\n",
      "==========================================================\n",
      "Epoch 11, Itrs 2010/10000, Loss=0.925970\n",
      "Epoch 11, Itrs 2020/10000, Loss=0.921045\n",
      "Epoch 11, Itrs 2030/10000, Loss=0.902406\n",
      "Epoch 11, Itrs 2040/10000, Loss=0.929929\n",
      "Epoch 11, Itrs 2050/10000, Loss=0.916792\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.06it/s]\n",
      "Validation Loss: 0.486106\n",
      "\n",
      "Foreground Acc: 0.963584\n",
      "Precision: 0.862000\n",
      "Recall: 0.963584\n",
      "F1 Score: 0.909966\n",
      "IoU Foreground: 0.834805\n",
      "IoU Background: 0.956218\n",
      "Mean IoU: 0.895511\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224, 0.8376933561199469, 0.8359486673807737, 0.8416792515202142, 0.8495187035894669, 0.8394195386637451, 0.8387325560736214, 0.8348048816371239]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700, 1750, 1800, 1850, 1900, 1950, 2000, 2050]\n",
      "==========================================================\n",
      "Epoch 11, Itrs 2060/10000, Loss=0.906814\n",
      "Epoch 11, Itrs 2070/10000, Loss=0.912777\n",
      "Epoch 11, Itrs 2080/10000, Loss=0.918307\n",
      "Epoch 11, Itrs 2090/10000, Loss=0.917106\n",
      "Epoch 12, Itrs 2100/10000, Loss=0.932175\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.24it/s]\n",
      "Validation Loss: 0.471653\n",
      "\n",
      "Foreground Acc: 0.956452\n",
      "Precision: 0.879265\n",
      "Recall: 0.956452\n",
      "F1 Score: 0.916236\n",
      "IoU Foreground: 0.845419\n",
      "IoU Background: 0.959910\n",
      "Mean IoU: 0.902665\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224, 0.8376933561199469, 0.8359486673807737, 0.8416792515202142, 0.8495187035894669, 0.8394195386637451, 0.8387325560736214, 0.8348048816371239, 0.8454194072776244]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700, 1750, 1800, 1850, 1900, 1950, 2000, 2050, 2100]\n",
      "==========================================================\n",
      "Epoch 12, Itrs 2110/10000, Loss=0.914385\n",
      "Epoch 12, Itrs 2120/10000, Loss=0.908431\n",
      "Epoch 12, Itrs 2130/10000, Loss=0.923842\n",
      "Epoch 12, Itrs 2140/10000, Loss=0.915147\n",
      "Epoch 12, Itrs 2150/10000, Loss=0.931934\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:11, 17.23it/s]\n",
      "Validation Loss: 0.479017\n",
      "\n",
      "Foreground Acc: 0.958192\n",
      "Precision: 0.877152\n",
      "Recall: 0.958192\n",
      "F1 Score: 0.915883\n",
      "IoU Foreground: 0.844819\n",
      "IoU Background: 0.959637\n",
      "Mean IoU: 0.902228\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224, 0.8376933561199469, 0.8359486673807737, 0.8416792515202142, 0.8495187035894669, 0.8394195386637451, 0.8387325560736214, 0.8348048816371239, 0.8454194072776244, 0.8448191245663028]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700, 1750, 1800, 1850, 1900, 1950, 2000, 2050, 2100, 2150]\n",
      "==========================================================\n",
      "Epoch 12, Itrs 2160/10000, Loss=0.917676\n",
      "Epoch 12, Itrs 2170/10000, Loss=0.932892\n",
      "Epoch 12, Itrs 2180/10000, Loss=0.917637\n",
      "Epoch 12, Itrs 2190/10000, Loss=0.910573\n",
      "Epoch 12, Itrs 2200/10000, Loss=0.900678\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:12, 16.95it/s]\n",
      "Validation Loss: 0.474932\n",
      "\n",
      "Foreground Acc: 0.959298\n",
      "Precision: 0.874995\n",
      "Recall: 0.959298\n",
      "F1 Score: 0.915209\n",
      "IoU Foreground: 0.843673\n",
      "IoU Background: 0.959226\n",
      "Mean IoU: 0.901450\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224, 0.8376933561199469, 0.8359486673807737, 0.8416792515202142, 0.8495187035894669, 0.8394195386637451, 0.8387325560736214, 0.8348048816371239, 0.8454194072776244, 0.8448191245663028, 0.8436732812052533]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700, 1750, 1800, 1850, 1900, 1950, 2000, 2050, 2100, 2150, 2200]\n",
      "==========================================================\n",
      "Epoch 12, Itrs 2210/10000, Loss=0.915378\n",
      "Epoch 12, Itrs 2220/10000, Loss=0.926379\n",
      "Epoch 12, Itrs 2230/10000, Loss=0.924768\n",
      "Epoch 12, Itrs 2240/10000, Loss=0.917069\n",
      "Epoch 12, Itrs 2250/10000, Loss=0.927444\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:12, 16.94it/s]\n",
      "Validation Loss: 0.487519\n",
      "\n",
      "Foreground Acc: 0.961841\n",
      "Precision: 0.869786\n",
      "Recall: 0.961841\n",
      "F1 Score: 0.913500\n",
      "IoU Foreground: 0.840773\n",
      "IoU Background: 0.958192\n",
      "Mean IoU: 0.899482\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224, 0.8376933561199469, 0.8359486673807737, 0.8416792515202142, 0.8495187035894669, 0.8394195386637451, 0.8387325560736214, 0.8348048816371239, 0.8454194072776244, 0.8448191245663028, 0.8436732812052533, 0.8407730717110262]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700, 1750, 1800, 1850, 1900, 1950, 2000, 2050, 2100, 2150, 2200, 2250]\n",
      "==========================================================\n",
      "Epoch 12, Itrs 2260/10000, Loss=0.905536\n",
      "Epoch 12, Itrs 2270/10000, Loss=0.927330\n",
      "Epoch 12, Itrs 2280/10000, Loss=0.922193\n",
      "Epoch 13, Itrs 2290/10000, Loss=0.912120\n",
      "Epoch 13, Itrs 2300/10000, Loss=0.918017\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "204it [00:12, 16.96it/s]\n",
      "Validation Loss: 0.477147\n",
      "\n",
      "Foreground Acc: 0.958681\n",
      "Precision: 0.877855\n",
      "Recall: 0.958681\n",
      "F1 Score: 0.916489\n",
      "IoU Foreground: 0.845852\n",
      "IoU Background: 0.959929\n",
      "Mean IoU: 0.902891\n",
      "\n",
      "[0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224, 0.8376933561199469, 0.8359486673807737, 0.8416792515202142, 0.8495187035894669, 0.8394195386637451, 0.8387325560736214, 0.8348048816371239, 0.8454194072776244, 0.8448191245663028, 0.8436732812052533, 0.8407730717110262, 0.8458518778740474]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700, 1750, 1800, 1850, 1900, 1950, 2000, 2050, 2100, 2150, 2200, 2250, 2300]\n",
      "==========================================================\n",
      "Epoch 13, Itrs 2310/10000, Loss=0.929353\n",
      "Epoch 13, Itrs 2320/10000, Loss=0.918171\n",
      "Epoch 13, Itrs 2330/10000, Loss=0.928558\n",
      "^C\n",
      "Traceback (most recent call last):\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/binary_leaf_main.py\", line 595, in <module>\n",
      "    main()\n",
      "  File \"/teamspace/studios/this_studio/Deeplab/binary_leaf_main.py\", line 523, in main\n",
      "    np_loss = loss.detach().cpu().numpy()\n",
      "KeyboardInterrupt\n"
     ]
    }
   ],
   "source": [
    "!python3 binary_leaf_main.py --model deeplabv3plus_mobilenet --custom_data_path \"india_sam_dino_annotations_large/\" --batch_size 16 --lr 0.01 --total_itrs 10000 --loss_type binary_dice --val_interval 50  --separable_conv\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "9a34fb58-687c-4a9b-afa2-52c41f697f8c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Example lists of x and y values\n",
    "x_values = [50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000, 1050, 1100, 1150, 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600, 1650, 1700, 1750, 1800, 1850, 1900, 1950, 2000, 2050, 2100, 2150, 2200, 2250, 2300]\n",
    "y_values = [0.4886676694697532, 0.6819128100575402, 0.7316776324091185, 0.7539703610134451, 0.7646733691379697, 0.7716763754696642, 0.780942755510218, 0.7891591265755157, 0.7947723249076609, 0.7976980095166158, 0.80284455982853, 0.7995927554839285, 0.8090232975237008, 0.8149058771937127, 0.8126718570911445, 0.8184854797204618, 0.8180831658173715, 0.8184463856799942, 0.8148966220170071, 0.8265067495056632, 0.8197575126821651, 0.8277434022634125, 0.8181608798921589, 0.8224712106517655, 0.8243492393484301, 0.8263397295553653, 0.8320330831656192, 0.8333233742814653, 0.8284184902499117, 0.8367651212779491, 0.8322137499776341, 0.8355160563935313, 0.8401478523183005, 0.8413897342847224, 0.8376933561199469, 0.8359486673807737, 0.8416792515202142, 0.8495187035894669, 0.8394195386637451, 0.8387325560736214, 0.8348048816371239, 0.8454194072776244, 0.8448191245663028, 0.8436732812052533, 0.8407730717110262, 0.8458518778740474]\n",
    "\n",
    "# Plot the graph\n",
    "plt.plot(x_values, y_values, marker='o', linestyle='-')\n",
    "plt.xlabel('Iterations')\n",
    "plt.ylabel('Foreground IOU')\n",
    "plt.title('India_Sam_Dino_Annotations_Training_large')\n",
    "plt.grid(True)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "59814251-f457-43f8-9e9c-040775054e8d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Example lists of x and y values\n",
    "x_values = [50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950]\n",
    "y_values = [0.77188253807654, 0.7945882804703195, 0.85439656783772, 0.8569383088344147, 0.8665472755014324, 0.8743193853513012, 0.8836554819011959, 0.8910747409996593, 0.8864961294929083, 0.8904289114374867, 0.8894175639405554, 0.8923611695996496, 0.8992382186585736, 0.9002104873381319, 0.8979438456833823, 0.9014707573662386, 0.9034829144092278, 0.9027885635096615, 0.8989919220038779]\n",
    "\n",
    "# Plot the graph\n",
    "plt.plot(x_values, y_values, marker='o', linestyle='-')\n",
    "plt.xlabel('Iterations')\n",
    "plt.ylabel('Foreground IOU')\n",
    "plt.title('India_Sam_Dino_Annotations_Training_small')\n",
    "plt.grid(True)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "7df2f572-f3aa-4366-8fa3-fed81926ee43",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device: cuda\n",
      "Dataset: custom, Train set: 1184, Val set: 315\n",
      "[!] Retrain\n",
      "Epoch 1, Itrs 10/10000, Loss=1.327529\n",
      "Epoch 1, Itrs 20/10000, Loss=1.173729\n",
      "Epoch 1, Itrs 30/10000, Loss=1.072247\n",
      "Epoch 1, Itrs 40/10000, Loss=1.026006\n",
      "Epoch 1, Itrs 50/10000, Loss=1.020204\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "79it [00:04, 16.83it/s]\n",
      "Validation Loss: 0.899830\n",
      "\n",
      "Foreground Acc: 0.319595\n",
      "Precision: 0.815647\n",
      "Recall: 0.319595\n",
      "F1 Score: 0.459244\n",
      "IoU Foreground: 0.298064\n",
      "IoU Background: 0.853826\n",
      "Mean IoU: 0.575945\n",
      "\n",
      "[0.29806400034784947]\n",
      "[50]\n",
      "Model saved as saved_models/100_percent_volume_inhouse_resized_0.29806400034784947.pth\n",
      "==========================================================\n",
      "Epoch 1, Itrs 60/10000, Loss=0.982481\n",
      "Epoch 1, Itrs 70/10000, Loss=0.943219\n",
      "Epoch 2, Itrs 80/10000, Loss=0.945691\n",
      "Epoch 2, Itrs 90/10000, Loss=0.936035\n",
      "Epoch 2, Itrs 100/10000, Loss=0.863783\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "79it [00:04, 16.12it/s]\n",
      "Validation Loss: 0.737878\n",
      "\n",
      "Foreground Acc: 0.517592\n",
      "Precision: 0.779805\n",
      "Recall: 0.517592\n",
      "F1 Score: 0.622201\n",
      "IoU Foreground: 0.451590\n",
      "IoU Background: 0.873042\n",
      "Mean IoU: 0.662316\n",
      "\n",
      "[0.29806400034784947, 0.4515904902529802]\n",
      "[50, 100]\n",
      "Model saved as saved_models/100_percent_volume_inhouse_resized_0.4515904902529802.pth\n",
      "==========================================================\n",
      "Epoch 2, Itrs 110/10000, Loss=0.903644\n",
      "Epoch 2, Itrs 120/10000, Loss=0.926524\n",
      "Epoch 2, Itrs 130/10000, Loss=0.871404\n",
      "Epoch 2, Itrs 140/10000, Loss=0.896398\n",
      "Epoch 3, Itrs 150/10000, Loss=0.912371\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "79it [00:04, 16.07it/s]\n",
      "Validation Loss: 0.711831\n",
      "\n",
      "Foreground Acc: 0.591546\n",
      "Precision: 0.739168\n",
      "Recall: 0.591546\n",
      "F1 Score: 0.657169\n",
      "IoU Foreground: 0.489391\n",
      "IoU Background: 0.873447\n",
      "Mean IoU: 0.681419\n",
      "\n",
      "[0.29806400034784947, 0.4515904902529802, 0.4893908143997243]\n",
      "[50, 100, 150]\n",
      "Model saved as saved_models/100_percent_volume_inhouse_resized_0.4893908143997243.pth\n",
      "==========================================================\n",
      "Epoch 3, Itrs 160/10000, Loss=0.789921\n",
      "Epoch 3, Itrs 170/10000, Loss=0.864982\n",
      "Epoch 3, Itrs 180/10000, Loss=0.834414\n",
      "Epoch 3, Itrs 190/10000, Loss=0.887330\n",
      "Epoch 3, Itrs 200/10000, Loss=0.831551\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "79it [00:04, 16.19it/s]\n",
      "Validation Loss: 0.743991\n",
      "\n",
      "Foreground Acc: 0.792545\n",
      "Precision: 0.609673\n",
      "Recall: 0.792545\n",
      "F1 Score: 0.689184\n",
      "IoU Foreground: 0.525767\n",
      "IoU Background: 0.847120\n",
      "Mean IoU: 0.686444\n",
      "\n",
      "[0.29806400034784947, 0.4515904902529802, 0.4893908143997243, 0.5257670810725178]\n",
      "[50, 100, 150, 200]\n",
      "Model saved as saved_models/100_percent_volume_inhouse_resized_0.5257670810725178.pth\n",
      "==========================================================\n",
      "Epoch 3, Itrs 210/10000, Loss=0.920113\n",
      "Epoch 3, Itrs 220/10000, Loss=0.807503\n",
      "Epoch 4, Itrs 230/10000, Loss=0.826905\n",
      "Epoch 4, Itrs 240/10000, Loss=0.837144\n",
      "Epoch 4, Itrs 250/10000, Loss=0.802295\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "79it [00:04, 16.03it/s]\n",
      "Validation Loss: 0.690598\n",
      "\n",
      "Foreground Acc: 0.722641\n",
      "Precision: 0.661025\n",
      "Recall: 0.722641\n",
      "F1 Score: 0.690461\n",
      "IoU Foreground: 0.527255\n",
      "IoU Background: 0.863475\n",
      "Mean IoU: 0.695365\n",
      "\n",
      "[0.29806400034784947, 0.4515904902529802, 0.4893908143997243, 0.5257670810725178, 0.5272552261617546]\n",
      "[50, 100, 150, 200, 250]\n",
      "Model saved as saved_models/100_percent_volume_inhouse_resized_0.5272552261617546.pth\n",
      "==========================================================\n",
      "Epoch 4, Itrs 260/10000, Loss=0.793679\n",
      "Epoch 4, Itrs 270/10000, Loss=0.778279\n",
      "Epoch 4, Itrs 280/10000, Loss=0.760041\n",
      "Epoch 4, Itrs 290/10000, Loss=0.782571\n",
      "Epoch 5, Itrs 300/10000, Loss=0.787915\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "79it [00:04, 16.05it/s]\n",
      "Validation Loss: 0.649636\n",
      "\n",
      "Foreground Acc: 0.644993\n",
      "Precision: 0.722994\n",
      "Recall: 0.644993\n",
      "F1 Score: 0.681770\n",
      "IoU Foreground: 0.517185\n",
      "IoU Background: 0.875168\n",
      "Mean IoU: 0.696177\n",
      "\n",
      "[0.29806400034784947, 0.4515904902529802, 0.4893908143997243, 0.5257670810725178, 0.5272552261617546, 0.5171854447078947]\n",
      "[50, 100, 150, 200, 250, 300]\n",
      "==========================================================\n",
      "Epoch 5, Itrs 310/10000, Loss=0.818030\n",
      "Epoch 5, Itrs 320/10000, Loss=0.737283\n",
      "Epoch 5, Itrs 330/10000, Loss=0.769860\n",
      "Epoch 5, Itrs 340/10000, Loss=0.728151\n",
      "Epoch 5, Itrs 350/10000, Loss=0.774937\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "79it [00:04, 16.14it/s]\n",
      "Validation Loss: 0.697736\n",
      "\n",
      "Foreground Acc: 0.750381\n",
      "Precision: 0.647788\n",
      "Recall: 0.750381\n",
      "F1 Score: 0.695320\n",
      "IoU Foreground: 0.532943\n",
      "IoU Background: 0.860620\n",
      "Mean IoU: 0.696782\n",
      "\n",
      "[0.29806400034784947, 0.4515904902529802, 0.4893908143997243, 0.5257670810725178, 0.5272552261617546, 0.5171854447078947, 0.5329433784118175]\n",
      "[50, 100, 150, 200, 250, 300, 350]\n",
      "Model saved as saved_models/100_percent_volume_inhouse_resized_0.5329433784118175.pth\n",
      "==========================================================\n",
      "Epoch 5, Itrs 360/10000, Loss=0.825713\n",
      "Epoch 5, Itrs 370/10000, Loss=0.759123\n",
      "Epoch 6, Itrs 380/10000, Loss=0.718147\n",
      "Epoch 6, Itrs 390/10000, Loss=0.720561\n",
      "Epoch 6, Itrs 400/10000, Loss=0.785162\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "79it [00:04, 16.13it/s]\n",
      "Validation Loss: 0.857499\n",
      "\n",
      "Foreground Acc: 0.848938\n",
      "Precision: 0.544782\n",
      "Recall: 0.848938\n",
      "F1 Score: 0.663671\n",
      "IoU Foreground: 0.496638\n",
      "IoU Background: 0.813742\n",
      "Mean IoU: 0.655190\n",
      "\n",
      "[0.29806400034784947, 0.4515904902529802, 0.4893908143997243, 0.5257670810725178, 0.5272552261617546, 0.5171854447078947, 0.5329433784118175, 0.4966377135747203]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400]\n",
      "==========================================================\n",
      "Epoch 6, Itrs 410/10000, Loss=0.763107\n",
      "Epoch 6, Itrs 420/10000, Loss=0.788490\n",
      "Epoch 6, Itrs 430/10000, Loss=0.728014\n",
      "Epoch 6, Itrs 440/10000, Loss=0.727569\n",
      "Epoch 7, Itrs 450/10000, Loss=0.753682\n",
      "Model saved as saved_models/latest_deeplabv3plus_mobilenet_custom_os16.pth\n",
      "validation...\n",
      "79it [00:04, 16.16it/s]\n",
      "Validation Loss: 0.662357\n",
      "\n",
      "Foreground Acc: 0.726625\n",
      "Precision: 0.682791\n",
      "Recall: 0.726625\n",
      "F1 Score: 0.704026\n",
      "IoU Foreground: 0.543241\n",
      "IoU Background: 0.871159\n",
      "Mean IoU: 0.707200\n",
      "\n",
      "[0.29806400034784947, 0.4515904902529802, 0.4893908143997243, 0.5257670810725178, 0.5272552261617546, 0.5171854447078947, 0.5329433784118175, 0.4966377135747203, 0.5432411918859545]\n",
      "[50, 100, 150, 200, 250, 300, 350, 400, 450]\n",
      "Model saved as saved_models/100_percent_volume_inhouse_resized_0.5432411918859545.pth\n",
      "==========================================================\n",
      "^C\n"
     ]
    }
   ],
   "source": [
    "!python3 binary_leaf_main.py --model deeplabv3plus_mobilenet --custom_data_path \"freetown_sam_dino_annotations/\" --batch_size 16 --lr 0.01 --total_itrs 10000 --loss_type binary_dice --val_interval 50  --separable_conv \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1a5797c1-7a8d-4140-a101-3b7c61f1da20",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c854e6f4-121f-4c43-bcd3-7fa3060c3827",
   "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": "3.10.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}