adv_patch_vanilla_bc.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/zeus/miniconda3/envs/cloudspace/lib/python3.9/site-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "# Python Notebook for Adversarial Patch Attack\n",
    "\n",
    "import os\n",
    "import hydra\n",
    "import torch\n",
    "from omegaconf import OmegaConf\n",
    "import pathlib\n",
    "from torch.utils.data import DataLoader\n",
    "import copy\n",
    "import random\n",
    "import dill\n",
    "import wandb\n",
    "import tqdm\n",
    "import numpy as np\n",
    "import shutil\n",
    "import collections\n",
    "from diffusion_policy.workspace.base_workspace import BaseWorkspace\n",
    "from diffusion_policy.policy.robomimic_image_policy import RobomimicImagePolicy\n",
    "from diffusion_policy.dataset.base_dataset import BaseImageDataset\n",
    "from diffusion_policy.dataset.robomimic_replay_image_dataset import RobomimicReplayImageDataset\n",
    "from diffusion_policy.env_runner.base_image_runner import BaseImageRunner\n",
    "from diffusion_policy.common.checkpoint_util import TopKCheckpointManager\n",
    "from diffusion_policy.common.json_logger import JsonLogger\n",
    "from diffusion_policy.common.pytorch_util import dict_apply, optimizer_to\n",
    "from diffusion_policy.workspace.train_robomimic_image_workspace import TrainRobomimicImageWorkspace\n",
    "from diffusion_policy.utils.attack_utils import transform_square_patch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import robomimic.utils.env_utils as EnvUtils\n",
    "import robomimic.utils.obs_utils as ObsUtils\n",
    "import robomimic.utils.file_utils as FileUtils\n",
    "\n",
    "def create_env(env_meta, shape_meta, enable_render=True):\n",
    "    modality_mapping = collections.defaultdict(list)\n",
    "    for key, attr in shape_meta['obs'].items():\n",
    "        modality_mapping[attr.get('type', 'low_dim')].append(key)\n",
    "    ObsUtils.initialize_obs_modality_mapping_from_dict(modality_mapping)\n",
    "    print(env_meta)\n",
    "\n",
    "    env = EnvUtils.create_env_from_metadata(\n",
    "        env_meta=env_meta,\n",
    "        render=False, \n",
    "        render_offscreen=enable_render,\n",
    "        use_image_obs=enable_render, \n",
    "    )\n",
    "    return env"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "============= Initialized Observation Utils with Obs Spec =============\n",
      "\n",
      "using obs modality: low_dim with keys: ['robot0_eef_quat', 'robot0_eef_pos', 'robot0_gripper_qpos']\n",
      "using obs modality: rgb with keys: ['robot0_eye_in_hand_image', 'agentview_image']\n",
      "using obs modality: depth with keys: []\n",
      "using obs modality: scan with keys: []\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/zeus/miniconda3/envs/cloudspace/lib/python3.9/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and will be removed in 0.15, please use 'weights' instead.\n",
      "  warnings.warn(\n",
      "/home/zeus/miniconda3/envs/cloudspace/lib/python3.9/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and will be removed in 0.15. The current behavior is equivalent to passing `weights=None`.\n",
      "  warnings.warn(msg)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'env_name': 'Lift', 'type': 1, 'env_kwargs': {'has_renderer': False, 'has_offscreen_renderer': True, 'ignore_done': True, 'use_object_obs': False, 'use_camera_obs': True, 'control_freq': 20, 'controller_configs': {'type': 'OSC_POSE', 'input_max': 1, 'input_min': -1, 'output_max': [0.05, 0.05, 0.05, 0.5, 0.5, 0.5], 'output_min': [-0.05, -0.05, -0.05, -0.5, -0.5, -0.5], 'kp': 150, 'damping': 1, 'impedance_mode': 'fixed', 'kp_limits': [0, 300], 'damping_limits': [0, 10], 'position_limits': None, 'orientation_limits': None, 'uncouple_pos_ori': True, 'control_delta': True, 'interpolation': None, 'ramp_ratio': 0.2}, 'robots': ['Panda'], 'camera_depths': False, 'camera_heights': 84, 'camera_widths': 84, 'reward_shaping': False, 'camera_names': ['agentview', 'robot0_eye_in_hand'], 'render_gpu_device_id': 0}}\n",
      "Created environment with name Lift\n",
      "Action size is 7\n"
     ]
    }
   ],
   "source": [
    "checkpoint = '/teamspace/studios/this_studio/bc_attacks/diffusion_policy/data/experiments/image/lift_ph/vanilla_bc/train_0/checkpoints/epoch=0500-test_mean_score=1.000.ckpt'\n",
    "payload = torch.load(open(checkpoint, 'rb'), pickle_module=dill)\n",
    "cfg_loaded = payload['cfg']\n",
    "cls = hydra.utils.get_class(cfg_loaded._target_)\n",
    "workspace = cls(cfg_loaded)\n",
    "workspace.load_payload(payload, exclude_keys=None, include_keys=None)\n",
    "policy = workspace.model\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "policy.to(device)\n",
    "policy.eval()\n",
    "\n",
    "cfg_loaded.task.env_runner['_target_'] = 'diffusion_policy.env_runner.robomimic_single_image_runner.RobomimicSingleImageRunner'\n",
    "cfg_loaded.task.env_runner.dataset_path = os.path.join('/teamspace/studios/this_studio/bc_attacks/diffusion_policy', cfg_loaded.task.dataset_path)\n",
    "env_runner = hydra.utils.instantiate(\n",
    "    cfg_loaded.task.env_runner,\n",
    "    output_dir=None\n",
    ")\n",
    "env = env_runner.env\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Timesteps: 100%|██████████| 100/100 [00:15<00:00,  6.50it/s]\n",
      "Timesteps: 100%|██████████| 100/100 [00:13<00:00,  7.30it/s]\n",
      "Timesteps: 100%|██████████| 100/100 [00:13<00:00,  7.57it/s]\n",
      "Timesteps: 100%|██████████| 100/100 [00:11<00:00,  8.42it/s]\n",
      "Timesteps:   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Timesteps:   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Timesteps:   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Timesteps:   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Timesteps:   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Timesteps:   0%|          | 0/100 [00:00<?, ?it/s]\n"
     ]
    }
   ],
   "source": [
    "# collect observations and actions for the policy for one rollout\n",
    "import tqdm\n",
    "obs = env.reset()\n",
    "obs_list = []\n",
    "action_list = []\n",
    "done = False\n",
    "timestep = 0\n",
    "max_timesteps = 100\n",
    "for i in range(10):\n",
    "    with tqdm.tqdm(total=max_timesteps, desc=\"Timesteps\") as pbar:\n",
    "        timestep = 0\n",
    "        while not done and timestep < max_timesteps:\n",
    "            np_obs_dict = dict(obs)\n",
    "            obs_dict = dict_apply(np_obs_dict, lambda x: torch.tensor(x).unsqueeze(0).to(device))\n",
    "            obs_list.append(obs_dict)\n",
    "            action_dict = policy.predict_action(obs_dict)\n",
    "            np_action_dict = dict_apply(action_dict, lambda x: x.cpu().detach().numpy())\n",
    "            action = np_action_dict['action'].squeeze(0)        \n",
    "            obs, _, done, _ = env.step(action)\n",
    "            action_list.append(action)\n",
    "            \n",
    "            timestep += 1\n",
    "            pbar.update(1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# squeeze the obs to shape (3, size, size)\n",
    "obs_list = [dict_apply(obs, lambda x: x.squeeze(0)) for obs in obs_list]\n",
    "# save the obs_list to load next time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x400 with 5 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualize few of the agentview_image observations\n",
    "import matplotlib.pyplot as plt\n",
    "from PIL import Image\n",
    "import numpy as np\n",
    "\n",
    "fig, axes = plt.subplots(1, 5, figsize=(20, 4))\n",
    "for i in range(5):\n",
    "    obs = obs_list[i]\n",
    "    agentview_image = obs['agentview_image'].squeeze(0).cpu().detach().numpy()\n",
    "    agentview_image = np.transpose(agentview_image, (1, 2, 0))\n",
    "    axes[i].imshow(agentview_image)\n",
    "    axes[i].axis('off')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# add perturbations to the actions\n",
    "perturbations = np.array([0.30, 0.30, 0., 0, 0, 0, 0])\n",
    "target_action_list = []\n",
    "for i in range(len(action_list)):\n",
    "    target_action = action_list[i] + perturbations\n",
    "    # target action to close the gripper always\n",
    "    # target_action[-1] = 1\n",
    "    target_action_list.append(target_action)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def place_patch(image, patch, location=(0, 0)):\n",
    "    # image: (B, n_obs, 3, size, size)\n",
    "    # patch: (3, patch_size, patch_size)\n",
    "    # location: (x, y)\n",
    "    original_len = len(image.shape)\n",
    "    if original_len == 4:\n",
    "        image = image.unsqueeze(0)\n",
    "    x, y = location\n",
    "    image[:, :, :, x:x+patch.shape[1], y:y+patch.shape[2]] = patch_forward(patch)\n",
    "    if original_len == 4:\n",
    "        image = image.squeeze(0)\n",
    "    return image\n",
    "\n",
    "def patch_forward(patch):\n",
    "    # applies tanh to the patch and scales it to [0, 1]\n",
    "    # patch = torch.tanh(patch) / 2 + 0.5\n",
    "    patch = 0.125 * torch.tanh(patch)\n",
    "    return patch\n",
    "\n",
    "def place_transparent_patch(image, patch, location=(0, 0)):\n",
    "    # image: (B, n_obs, 3, size, size)\n",
    "    # patch: (3, patch_size, patch_size)\n",
    "    # location: (x, y)\n",
    "    original_len = len(image.shape)\n",
    "    if original_len == 4:\n",
    "        image = image.unsqueeze(0)\n",
    "    x, y = location\n",
    "    patch = patch_forward(patch)\n",
    "    # image[:, :, :, x:x+patch.shape[1], y:y+patch.shape[2]] = patch_alpha * patch[:3] + (1 - patch_alpha) * image[:, :, :, x:x+patch.shape[1], y:y+patch.shape[2]]\n",
    "    image[:, :, :, :, :] = image[:, :, :, :, :] + patch[:3]\n",
    "    if original_len == 4:\n",
    "        image = image.squeeze(0)\n",
    "    return image\n",
    "\n",
    "# visualize the patch\n",
    "patch_ = torch.zeros((3, 32, 32))\n",
    "image_ = obs_list[0]['robot0_eye_in_hand_image'].clone()\n",
    "image_ = place_patch(image_, patch_, location=(24, 24))\n",
    "image_ = image_.squeeze(0).cpu().detach().numpy()\n",
    "image_ = np.transpose(image_, (1, 2, 0))\n",
    "plt.imshow(image_)\n",
    "plt.axis('off')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_loc, y_loc = 16, 16\n",
    "patch_size = 16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# evaluate the patch\n",
    "view = 'robot0_eye_in_hand_image'\n",
    "# view = 'agentview_image'\n",
    "def evaluate_patch(patch, transparent=False):\n",
    "    patch_cp = patch.clone().detach()\n",
    "    obs = env.reset()\n",
    "    policy.reset()\n",
    "    obs = dict_apply(obs, lambda x: torch.tensor(x).to(device))\n",
    "    if transparent:\n",
    "        obs[view] = place_transparent_patch(obs[view], patch_cp, location=(0, 0))\n",
    "    else:\n",
    "        obs[view] = place_patch(obs[view], patch_cp, location=(x_loc, y_loc))\n",
    "    # convert to tensor\n",
    "    obs = dict_apply(obs, lambda x: torch.tensor(x).unsqueeze(0))\n",
    "    done = False\n",
    "    max_timesteps = 400\n",
    "    rewards = 0\n",
    "    obs_agentview = []\n",
    "    obs_robot0_eye_in_hand = []\n",
    "    with tqdm.tqdm(total=max_timesteps, desc=\"Timesteps\") as pbar:\n",
    "        timestep = 0\n",
    "        while not done and timestep < max_timesteps:\n",
    "            # for key in obs.keys():\n",
    "            #    print(key, obs[key].shape)\n",
    "            action_dict = policy.predict_action(obs)\n",
    "            action = action_dict['action'].cpu().detach().numpy().squeeze(0)\n",
    "            obs, reward, done, _ = env.step(action)\n",
    "            obs = dict_apply(obs, lambda x: torch.tensor(x).to(device))\n",
    "            if transparent:\n",
    "                obs[view] = place_transparent_patch(obs[view], patch_cp, location=(0, 0))\n",
    "            else:\n",
    "                obs[view] = place_patch(obs[view], patch_cp, location=(x_loc, y_loc))\n",
    "            # obs = dict_apply(obs, lambda x: x.squeeze(0))\n",
    "            obs = dict_apply(obs, lambda x: torch.tensor(x).unsqueeze(0))\n",
    "            obs_agentview.append(obs['agentview_image'])\n",
    "            obs_robot0_eye_in_hand.append(obs['robot0_eye_in_hand_image'])\n",
    "            timestep += 1\n",
    "            pbar.update(1)\n",
    "            rewards += reward\n",
    "        if rewards > 0:\n",
    "            print(f\"Patch is unsuccessful with reward: {rewards}\")\n",
    "        elif rewards == 0:\n",
    "            print(f\"Patch is successful with reward: {rewards}\")\n",
    "        else:\n",
    "            pass\n",
    "    return rewards, obs_agentview, obs_robot0_eye_in_hand\n",
    "\n",
    "# test the evaluate_patch function\n",
    "# patch = torch.zeros((3, patch_size, patch_size))\n",
    "# red_patch = torch.zeros((3, patch_size, patch_size))\n",
    "# red_patch[0] = 1\n",
    "# rewards, observation_agentview, observation_robot0_eye_in_hand = evaluate_patch(red_patch)\n",
    "# print(rewards)\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "from PIL import Image\n",
    "import numpy as np\n",
    "from matplotlib.animation import FuncAnimation, ArtistAnimation\n",
    "\n",
    "\n",
    "def visualize(observations1, observations2, file_name, every_frame=2):\n",
    "    # only get alternate frames to reduce the processing time\n",
    "    observations1 = observations1[::every_frame]\n",
    "    observations2 = observations2[::every_frame]\n",
    "    im = []\n",
    "    for i in range(len(observations1)):\n",
    "        agentview_image1 = observations1[i].squeeze(0).squeeze(0).cpu().detach().numpy()\n",
    "        agentview_image1 = np.transpose(agentview_image1, (1, 2, 0))\n",
    "        \n",
    "        agentview_image2 = observations2[i].squeeze(0).squeeze(0).cpu().detach().numpy()\n",
    "        agentview_image2 = np.transpose(agentview_image2, (1, 2, 0))\n",
    "        \n",
    "        combined_image = np.hstack((agentview_image1, agentview_image2))\n",
    "        im.append(combined_image)\n",
    "\n",
    "    fig, ax = plt.subplots()\n",
    "    ims = []\n",
    "    for i in range(len(im)):\n",
    "        im_obj = ax.imshow(im[i], animated=True)\n",
    "        ims.append([im_obj])\n",
    "        ax.axis('off')\n",
    "\n",
    "    ani = ArtistAnimation(fig, ims, interval=100, blit=True)\n",
    "    ani.save(file_name, writer='pillow', fps=10)\n",
    "\n",
    "# visualize the observations\n",
    "# visualize(observation_robot0_eye_in_hand, observation_agentview, file_name = f'/teamspace/studios/this_studio/bc_attacks/diffusion_policy/plots/videos/adversarial_patch/red_patch_x_{x_loc}_y_{y_loc}_{patch_size}_rew_{rewards}.gif')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch.utils.data import Dataset\n",
    "class ObservationActionDataset(Dataset):\n",
    "    def __init__(self, observations, actions):\n",
    "        self.observations = observations\n",
    "        self.actions = actions\n",
    "\n",
    "    def __len__(self):\n",
    "        return len(self.observations)\n",
    "\n",
    "    def __getitem__(self, idx):\n",
    "        observation = self.observations[idx]\n",
    "        action = self.actions[idx]\n",
    "        return observation, action"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Iterations:   0%|          | 0/200 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Iterations:   0%|          | 0/200 [00:00<?, ?it/s]\n"
     ]
    },
    {
     "ename": "NameError",
     "evalue": "name 'action_means' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[27], line 61\u001b[0m\n\u001b[1;32m     59\u001b[0m patch_sizes \u001b[38;5;241m=\u001b[39m [\u001b[38;5;241m24\u001b[39m]\n\u001b[1;32m     60\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m patch_size \u001b[38;5;129;01min\u001b[39;00m patch_sizes:\n\u001b[0;32m---> 61\u001b[0m     patch \u001b[38;5;241m=\u001b[39m \u001b[43mpatch_attack\u001b[49m\u001b[43m(\u001b[49m\u001b[43mpolicy\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mtrain_loader\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mval_loader\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mpatch_size\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mnum_iterations\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;241;43m200\u001b[39;49m\u001b[43m)\u001b[49m\n\u001b[1;32m     62\u001b[0m     \u001b[38;5;66;03m# evaluate the patch\u001b[39;00m\n\u001b[1;32m     63\u001b[0m     rewards, observation_patch_agentview, observation_patch_robot0 \u001b[38;5;241m=\u001b[39m evaluate_patch(patch)\n",
      "Cell \u001b[0;32mIn[27], line 20\u001b[0m, in \u001b[0;36mpatch_attack\u001b[0;34m(policy, train_loader, val_loader, patch_size, num_iterations)\u001b[0m\n\u001b[1;32m     18\u001b[0m obs[view] \u001b[38;5;241m=\u001b[39m place_patch(obs[view], patch, location\u001b[38;5;241m=\u001b[39m(x_loc, y_loc))\n\u001b[1;32m     19\u001b[0m actions \u001b[38;5;241m=\u001b[39m policy\u001b[38;5;241m.\u001b[39mpredict_action(obs)[\u001b[38;5;124m'\u001b[39m\u001b[38;5;124maction\u001b[39m\u001b[38;5;124m'\u001b[39m]\n\u001b[0;32m---> 20\u001b[0m loss \u001b[38;5;241m=\u001b[39m torch\u001b[38;5;241m.\u001b[39mnn\u001b[38;5;241m.\u001b[39mfunctional\u001b[38;5;241m.\u001b[39mmse_loss(\u001b[43maction_means\u001b[49m, torch\u001b[38;5;241m.\u001b[39mtensor(target_action, dtype\u001b[38;5;241m=\u001b[39mtorch\u001b[38;5;241m.\u001b[39mfloat32)\u001b[38;5;241m.\u001b[39mto(device))\n\u001b[1;32m     21\u001b[0m optimizer\u001b[38;5;241m.\u001b[39mzero_grad()\n\u001b[1;32m     22\u001b[0m loss\u001b[38;5;241m.\u001b[39mbackward()\n",
      "\u001b[0;31mNameError\u001b[0m: name 'action_means' is not defined"
     ]
    }
   ],
   "source": [
    "# disable cudann\n",
    "torch.backends.cudnn.enabled = False\n",
    "# create a patch of size 24x24\n",
    "def patch_attack(policy, train_loader, val_loader, patch_size, num_iterations=100):\n",
    "    # create a patch of size 24x24\n",
    "    patch = torch.zeros((3, patch_size, patch_size), dtype=torch.float32)\n",
    "    clean_patch = patch.clone().detach().to(device)\n",
    "    patch = patch.to(device)\n",
    "    # convert the patch to a parameter which can be optimized\n",
    "    patch = torch.nn.Parameter(patch)\n",
    "    optimizer = torch.optim.Adam([patch], lr=0.01)\n",
    "    losses = []\n",
    "    with tqdm.tqdm(total=num_iterations, desc=\"Iterations\") as pbar:\n",
    "        for i in range(num_iterations):\n",
    "            total_loss = 0\n",
    "            for obs, target_action in train_loader:\n",
    "                policy.reset()\n",
    "                obs[view] = place_patch(obs[view], patch, location=(x_loc, y_loc))\n",
    "                actions = policy.predict_action(obs)['action']\n",
    "                loss = torch.nn.functional.mse_loss(action_means, torch.tensor(target_action, dtype=torch.float32).to(device))\n",
    "                optimizer.zero_grad()\n",
    "                loss.backward()\n",
    "                optimizer.step()\n",
    "                total_loss += loss.item()\n",
    "            losses.append(total_loss)\n",
    "            # print(f\"Change in patch: {torch.norm(patch - clean_patch)}\")\n",
    "            if i % 100 == 0 and i != 0:\n",
    "                print(f\"Iteration: {i}, Loss: {total_loss}\")\n",
    "                # evaluate_patch(patch)\n",
    "                # visualize the patch\n",
    "                patch_np = patch.cpu().detach().numpy()\n",
    "                patch_np = np.transpose(patch_np, (1, 2, 0))\n",
    "                plt.imshow(patch_np)\n",
    "                plt.axis('off')\n",
    "                plt.show()\n",
    "            pbar.update(1)\n",
    "    # plot the losses\n",
    "    plt.plot(losses)\n",
    "    plt.xlabel('Iterations')\n",
    "    plt.ylabel('Loss')\n",
    "    plt.title('Loss vs Iterations')\n",
    "    plt.show()\n",
    "    return patch\n",
    "\n",
    "# convert the obs_list and target_action_list to tensors as a dataset\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "from sklearn.model_selection import train_test_split\n",
    "# Split data into training and validation sets\n",
    "obs_train, obs_val, action_train, action_val = train_test_split(obs_list, target_action_list, test_size=0.2, random_state=42)\n",
    "# Create Dataset instances\n",
    "train_dataset = ObservationActionDataset(obs_train, action_train)\n",
    "val_dataset = ObservationActionDataset(obs_val, action_val)\n",
    "\n",
    "# Create DataLoaders\n",
    "train_loader = DataLoader(train_dataset, batch_size=32, shuffle=True)\n",
    "val_loader = DataLoader(val_dataset, batch_size=32, shuffle=False)\n",
    "\n",
    "# patch_sizes = [20, 24, 28, 32]\n",
    "patch_sizes = [24]\n",
    "for patch_size in patch_sizes:\n",
    "    patch = patch_attack(policy, train_loader, val_loader, patch_size, num_iterations=200)\n",
    "    # evaluate the patch\n",
    "    rewards, observation_patch_agentview, observation_patch_robot0 = evaluate_patch(patch)\n",
    "    print(f\"Reward: {rewards}\")\n",
    "    # visualize the observations\n",
    "    file_name = f'/teamspace/studios/this_studio/bc_attacks/diffusion_policy/plots/videos/adversarial_patch/trained_patch_perturbed_y15_x_{x_loc}_y_{y_loc}_{patch_size}_rew_{rewards}.gif'\n",
    "    visualize(observation_patch_robot0, observation_patch_agentview, file_name = file_name)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Iterations:   0%|          | 0/400 [00:00<?, ?it/s]/tmp/ipykernel_7303/2105181730.py:26: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n",
      "  loss = torch.nn.functional.mse_loss(actions, torch.tensor(target_action, dtype=torch.float32).to(device))\n",
      "Iterations:  25%|██▌       | 100/400 [01:07<03:23,  1.48it/s]WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration: 100, Loss: 0.17718549724668264\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Iterations:  33%|███▎      | 133/400 [01:29<03:03,  1.46it/s]"
     ]
    }
   ],
   "source": [
    "# disable cudann\n",
    "x_loc_trans, y_loc_trans = 0, 0\n",
    "torch.backends.cudnn.enabled = False\n",
    "epsilon = 0.125\n",
    "# create a patch of size 24x24\n",
    "def patch_attack(policy, train_loader, val_loader, patch_size, num_iterations=100):\n",
    "    # create a universal patch of size 84x84\n",
    "    universal_patch = torch.zeros((3, patch_size, patch_size), dtype=torch.float32)\n",
    "    # clean_patch = patch.clone().detach().to(device)\n",
    "    # patch = patch.to(device)\n",
    "    # convert the patch to a parameter which can be optimized\n",
    "    # patch = torch.nn.Parameter(patch)\n",
    "    universal_patch = universal_patch.to(device)\n",
    "    # convert the patch to a parameter which can be optimized\n",
    "    universal_patch = torch.nn.Parameter(universal_patch)\n",
    "    optimizer = torch.optim.Adam([universal_patch], lr=0.005)\n",
    "    losses = []\n",
    "    with tqdm.tqdm(total=num_iterations, desc=\"Iterations\") as pbar:\n",
    "        for i in range(num_iterations):\n",
    "            total_loss = 0\n",
    "            for obs, target_action in train_loader:\n",
    "                policy.reset()\n",
    "                obs[view] = place_transparent_patch(obs[view], universal_patch, location=(x_loc_trans, y_loc_trans))\n",
    "                obs[view] = torch.clamp(obs[view], 0, 1)\n",
    "                actions = policy.predict_action(obs)['action']\n",
    "                loss = torch.nn.functional.mse_loss(actions, torch.tensor(target_action, dtype=torch.float32).to(device))\n",
    "                optimizer.zero_grad()\n",
    "                loss.backward()\n",
    "                optimizer.step()\n",
    "                # clip the universal patch to be within the epsilon ball\n",
    "                # with torch.no_grad():\n",
    "                #    universal_patch = torch.clamp(universal_patch, -epsilon, epsilon)\n",
    "                # clip the observation to be within [0, 1]\n",
    "                total_loss += loss.item()\n",
    "            losses.append(total_loss)\n",
    "            # print(universal_patch.min(), universal_patch.max())\n",
    "            # print(f\"Change in patch: {torch.norm(patch - clean_patch)}\")\n",
    "            if i % 100 == 0 and i != 0:\n",
    "                print(f\"Iteration: {i}, Loss: {total_loss}\")\n",
    "                # evaluate_patch(patch)\n",
    "                # visualize the patch\n",
    "                patch_np = universal_patch.cpu().detach().numpy()\n",
    "                patch_np = np.transpose(patch_np, (1, 2, 0))\n",
    "                plt.imshow(patch_np)\n",
    "                plt.axis('off')\n",
    "                plt.show()\n",
    "            pbar.update(1)\n",
    "    # plot the losses\n",
    "    plt.plot(losses)\n",
    "    plt.xlabel('Iterations')\n",
    "    plt.ylabel('Loss')\n",
    "    plt.title('Loss vs Iterations')\n",
    "    plt.show()\n",
    "    return universal_patch\n",
    "\n",
    "# convert the obs_list and target_action_list to tensors as a dataset\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "from sklearn.model_selection import train_test_split\n",
    "# Split data into training and validation sets\n",
    "obs_train, obs_val, action_train, action_val = train_test_split(obs_list, target_action_list, test_size=0.2, random_state=42)\n",
    "# Create Dataset instances\n",
    "train_dataset = ObservationActionDataset(obs_train, action_train)\n",
    "val_dataset = ObservationActionDataset(obs_val, action_val)\n",
    "\n",
    "# Create DataLoaders\n",
    "train_loader = DataLoader(train_dataset, batch_size=32, shuffle=True)\n",
    "val_loader = DataLoader(val_dataset, batch_size=32, shuffle=False)\n",
    "\n",
    "# patch_sizes = [20, 24, 28, 32]\n",
    "patch_sizes = [84]\n",
    "for patch_size in patch_sizes:\n",
    "    universal_patch = patch_attack(policy, train_loader, val_loader, patch_size, num_iterations=400)\n",
    "    print(universal_patch.min(), universal_patch.max())\n",
    "    # save the universal_patch as a numpy file\n",
    "    universal_patch_np = universal_patch.cpu().detach().numpy()\n",
    "    np.save(f'/teamspace/studios/this_studio/bc_attacks/diffusion_policy/plots/universal_patch/vanilla_bc_universal_patch_transparent_xy_30_400iters_eps_0.125.npy', universal_patch_np)\n",
    "    # evaluate the patch\n",
    "    rewards, observation_patch_agentview, observation_patch_robot0 = evaluate_patch(universal_patch, transparent=True)\n",
    "    print(f\"Reward: {rewards}\")\n",
    "    # visualize the observations\n",
    "    file_name = f'/teamspace/studios/this_studio/bc_attacks/diffusion_policy/plots/videos/adversarial_patch/vanilla_bc_trained_transparent_patch_perturbed_x_30_y30_rew_{rewards}_eps0.125.gif'\n",
    "    visualize(observation_patch_robot0, observation_patch_agentview, file_name = file_name)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "============= Initialized Observation Utils with Obs Spec =============\n",
      "\n",
      "using obs modality: low_dim with keys: ['robot0_gripper_qpos', 'robot0_eef_quat', 'robot0_eef_pos']\n",
      "using obs modality: rgb with keys: ['agentview_image', 'robot0_eye_in_hand_image']\n",
      "using obs modality: depth with keys: []\n",
      "using obs modality: scan with keys: []\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/zeus/miniconda3/envs/cloudspace/lib/python3.9/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and will be removed in 0.15, please use 'weights' instead.\n",
      "  warnings.warn(\n",
      "/home/zeus/miniconda3/envs/cloudspace/lib/python3.9/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and will be removed in 0.15. The current behavior is equivalent to passing `weights=None`.\n",
      "  warnings.warn(msg)\n",
      "/tmp/ipykernel_5782/3492069373.py:14: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n",
      "  obs = dict_apply(obs, lambda x: torch.tensor(x).unsqueeze(0))\n",
      "Timesteps:   0%|          | 0/400 [00:00<?, ?it/s]/tmp/ipykernel_5782/3492069373.py:34: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n",
      "  obs = dict_apply(obs, lambda x: torch.tensor(x).unsqueeze(0))\n",
      "Timesteps: 100%|██████████| 400/400 [00:54<00:00,  7.37it/s]\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Patch is successful with reward: 0.0\n",
      "Reward: 0.0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# load other checkpoints policy to check the performance of the patch\n",
    "checkpoint = '/teamspace/studios/this_studio/bc_attacks/diffusion_policy/data/experiments/image/lift_ph/vanilla_bc/train_0/checkpoints/epoch=0500-test_mean_score=1.000.ckpt'\n",
    "payload = torch.load(open(checkpoint, 'rb'), pickle_module=dill)\n",
    "cfg_loaded = payload['cfg']\n",
    "cls = hydra.utils.get_class(cfg_loaded._target_)\n",
    "workspace = cls(cfg_loaded)\n",
    "workspace.load_payload(payload, exclude_keys=None, include_keys=None)\n",
    "policy = workspace.model\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "policy.to(device)\n",
    "policy.eval()\n",
    "\n",
    "# evaluate the patch\n",
    "rewards, observation_patch_agentview, observation_patch_robot0 = evaluate_patch(universal_patch, transparent=True)\n",
    "print(f\"Reward: {rewards}\")\n",
    "# visualize the observations\n",
    "file_name = f'/teamspace/studios/this_studio/bc_attacks/diffusion_policy/plots/videos/adversarial_patch/vanilla_bc_trained0_transparent_patch_perturbed_y15_rew_{rewards}.gif'\n",
    "visualize(observation_patch_robot0, observation_patch_agentview, file_name = file_name)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor(-3.0544, device='cuda:0', grad_fn=<MinBackward1>) tensor(2.9877, device='cuda:0', grad_fn=<MaxBackward1>)\n"
     ]
    }
   ],
   "source": [
    "print(universal_patch.min(), universal_patch.max())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_31751/878138627.py:14: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n",
      "  obs = dict_apply(obs, lambda x: torch.tensor(x).unsqueeze(0))\n",
      "Timesteps:   0%|          | 0/400 [00:00<?, ?it/s]/tmp/ipykernel_31751/878138627.py:34: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n",
      "  obs = dict_apply(obs, lambda x: torch.tensor(x).unsqueeze(0))\n",
      "Timesteps: 100%|██████████| 400/400 [00:53<00:00,  7.46it/s]\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Patch is successful with reward: 0.0\n",
      "Reward: 0.0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "patch_size = 84\n",
    "rewards, observation_patch_agentview, observation_patch_robot0 = evaluate_patch(universal_patch, transparent=True)\n",
    "print(f\"Reward: {rewards}\")\n",
    "# visualize the observations\n",
    "file_name = f'/teamspace/studios/this_studio/bc_attacks/diffusion_policy/plots/videos/adversarial_patch/trained_transparent_patch_perturbed_y60_x_{x_loc_trans}_y_{y_loc_trans}_{patch_size}_rew_{rewards}.gif'\n",
    "visualize(observation_patch_robot0, observation_patch_agentview, file_name = file_name)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'dict' object has no attribute 'shape'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[31], line 3\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[39m# visualize how the patch looks like on the robot0_eye_in_hand_image\u001b[39;00m\n\u001b[1;32m      2\u001b[0m obs_robot0 \u001b[39m=\u001b[39m obs_list[\u001b[39m0\u001b[39m]\n\u001b[0;32m----> 3\u001b[0m obs_robot0 \u001b[39m=\u001b[39m place_patch(obs_robot0, patch_24, location\u001b[39m=\u001b[39;49m(\u001b[39m16\u001b[39;49m, \u001b[39m16\u001b[39;49m))\n\u001b[1;32m      4\u001b[0m obs_robot0 \u001b[39m=\u001b[39m obs_robot0\u001b[39m.\u001b[39msqueeze(\u001b[39m0\u001b[39m)\u001b[39m.\u001b[39msqueeze(\u001b[39m0\u001b[39m)\u001b[39m.\u001b[39mcpu()\u001b[39m.\u001b[39mdetach()\u001b[39m.\u001b[39mnumpy()\n\u001b[1;32m      5\u001b[0m obs_robot0 \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39mtranspose(obs_robot0, (\u001b[39m1\u001b[39m, \u001b[39m2\u001b[39m, \u001b[39m0\u001b[39m))\n",
      "Cell \u001b[0;32mIn[25], line 5\u001b[0m, in \u001b[0;36mplace_patch\u001b[0;34m(image, patch, location)\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[39mdef\u001b[39;00m \u001b[39mplace_patch\u001b[39m(image, patch, location\u001b[39m=\u001b[39m(\u001b[39m0\u001b[39m, \u001b[39m0\u001b[39m)):\n\u001b[1;32m      2\u001b[0m     \u001b[39m# image: (B, n_obs, 3, size, size)\u001b[39;00m\n\u001b[1;32m      3\u001b[0m     \u001b[39m# patch: (3, patch_size, patch_size)\u001b[39;00m\n\u001b[1;32m      4\u001b[0m     \u001b[39m# location: (x, y)\u001b[39;00m\n\u001b[0;32m----> 5\u001b[0m     original_len \u001b[39m=\u001b[39m \u001b[39mlen\u001b[39m(image\u001b[39m.\u001b[39;49mshape)\n\u001b[1;32m      6\u001b[0m     \u001b[39mif\u001b[39;00m original_len \u001b[39m==\u001b[39m \u001b[39m4\u001b[39m:\n\u001b[1;32m      7\u001b[0m         image \u001b[39m=\u001b[39m image\u001b[39m.\u001b[39munsqueeze(\u001b[39m0\u001b[39m)\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'dict' object has no attribute 'shape'"
     ]
    }
   ],
   "source": [
    "# visualize how the patch looks like on the robot0_eye_in_hand_image\n",
    "obs_robot0 = obs_list[0]\n",
    "obs_robot0 = place_patch(obs_robot0, patch, location=(16, 16))\n",
    "obs_robot0 = obs_robot0.squeeze(0).squeeze(0).cpu().detach().numpy()\n",
    "obs_robot0 = np.transpose(obs_robot0, (1, 2, 0))\n",
    "plt.imshow(obs_robot0)\n",
    "plt.axis('off')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.18"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}