Spaces:
Runtime error

runtime error

Space not ready. Reason: Completed, exitCode: 0, message: None

Container logs:

2022-05-10 15:46:58.535574: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcudart.so.11.0'; dlerror: libcudart.so.11.0: cannot open shared object file: No such file or directory
2022-05-10 15:46:58.535722: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.

Downloading:   0%|          | 0.00/0.99M [00:00<?, ?B/s]
Downloading: 100%|██████████| 0.99M/0.99M [00:00<00:00, 54.0MB/s]

Downloading:   0%|          | 0.00/446k [00:00<?, ?B/s]
Downloading: 100%|██████████| 446k/446k [00:00<00:00, 40.7MB/s]

Downloading:   0%|          | 0.00/665 [00:00<?, ?B/s]
Downloading: 100%|██████████| 665/665 [00:00<00:00, 694kB/s]

Downloading:   0%|          | 0.00/523M [00:00<?, ?B/s]
Downloading:   1%|▏         | 7.47M/523M [00:00<00:06, 78.4MB/s]
Downloading:   3%|▎         | 15.8M/523M [00:00<00:06, 83.7MB/s]
Downloading:   5%|▍         | 24.1M/523M [00:00<00:06, 85.3MB/s]
Downloading:   6%|▌         | 32.5M/523M [00:00<00:05, 86.2MB/s]
Downloading:   8%|▊         | 40.8M/523M [00:00<00:05, 86.6MB/s]
Downloading:   9%|▉         | 49.2M/523M [00:00<00:05, 87.0MB/s]
Downloading:  11%|█         | 57.6M/523M [00:00<00:05, 87.3MB/s]
Downloading:  13%|█▎        | 66.0M/523M [00:00<00:05, 87.7MB/s]
Downloading:  14%|█▍        | 74.4M/523M [00:00<00:05, 87.6MB/s]
Downloading:  16%|█▌        | 82.8M/523M [00:01<00:05, 87.7MB/s]
Downloading:  17%|█▋        | 91.1M/523M [00:01<00:05, 87.7MB/s]
Downloading:  19%|█▉        | 99.5M/523M [00:01<00:05, 87.7MB/s]
Downloading:  21%|██        | 108M/523M [00:01<00:04, 87.7MB/s] 
Downloading:  22%|██▏       | 116M/523M [00:01<00:04, 87.6MB/s]
Downloading:  24%|██▍       | 125M/523M [00:01<00:04, 87.5MB/s]
Downloading:  25%|██▌       | 133M/523M [00:01<00:04, 87.7MB/s]
Downloading:  27%|██▋       | 141M/523M [00:01<00:04, 87.5MB/s]
Downloading:  29%|██▊       | 150M/523M [00:01<00:04, 87.8MB/s]
Downloading:  30%|███       | 158M/523M [00:01<00:04, 87.6MB/s]
Downloading:  32%|███▏      | 167M/523M [00:02<00:04, 87.7MB/s]
Downloading:  33%|███▎      | 175M/523M [00:02<00:04, 87.8MB/s]
Downloading:  35%|███▌      | 183M/523M [00:02<00:04, 87.7MB/s]
Downloading:  37%|███▋      | 192M/523M [00:02<00:03, 87.5MB/s]
Downloading:  38%|███▊      | 200M/523M [00:02<00:03, 87.0MB/s]
Downloading:  40%|███▉      | 208M/523M [00:02<00:03, 87.1MB/s]
Downloading:  41%|████▏     | 217M/523M [00:02<00:03, 87.3MB/s]
Downloading:  43%|████▎     | 225M/523M [00:02<00:03, 87.5MB/s]
Downloading:  45%|████▍     | 234M/523M [00:02<00:03, 87.8MB/s]
Downloading:  46%|████▋     | 242M/523M [00:02<00:03, 87.7MB/s]
Downloading:  48%|████▊     | 250M/523M [00:03<00:03, 87.7MB/s]
Downloading:  49%|████▉     | 259M/523M [00:03<00:03, 87.8MB/s]
Downloading:  51%|█████     | 267M/523M [00:03<00:03, 87.7MB/s]
Downloading:  53%|█████▎    | 275M/523M [00:03<00:02, 87.7MB/s]
Downloading:  54%|█████▍    | 284M/523M [00:03<00:02, 87.6MB/s]
Downloading:  56%|█████▌    | 292M/523M [00:03<00:02, 87.5MB/s]
Downloading:  57%|█████▋    | 301M/523M [00:03<00:02, 87.5MB/s]
Downloading:  59%|█████▉    | 309M/523M [00:03<00:02, 87.6MB/s]
Downloading:  61%|██████    | 317M/523M [00:03<00:02, 87.6MB/s]
Downloading:  62%|██████▏   | 326M/523M [00:03<00:02, 87.6MB/s]
Downloading:  64%|██████▍   | 334M/523M [00:04<00:02, 87.5MB/s]
Downloading:  65%|██████▌   | 342M/523M [00:04<00:02, 87.5MB/s]
Downloading:  67%|██████▋   | 351M/523M [00:04<00:02, 87.6MB/s]
Downloading:  69%|██████▊   | 359M/523M [00:04<00:01, 87.7MB/s]
Downloading:  70%|███████   | 367M/523M [00:04<00:01, 87.8MB/s]
Downloading:  72%|███████▏  | 376M/523M [00:04<00:01, 87.8MB/s]
Downloading:  74%|███████▎  | 384M/523M [00:04<00:01, 87.8MB/s]
Downloading:  75%|███████▌  | 393M/523M [00:04<00:01, 87.9MB/s]
Downloading:  77%|███████▋  | 401M/523M [00:04<00:01, 87.9MB/s]
Downloading:  78%|███████▊  | 409M/523M [00:04<00:01, 87.8MB/s]
Downloading:  80%|███████▉  | 418M/523M [00:05<00:01, 87.4MB/s]
Downloading:  82%|████████▏ | 426M/523M [00:05<00:01, 86.7MB/s]
Downloading:  83%|████████▎ | 434M/523M [00:05<00:01, 86.6MB/s]
Downloading:  85%|████████▍ | 443M/523M [00:05<00:00, 86.6MB/s]
Downloading:  86%|████████▋ | 451M/523M [00:05<00:00, 86.8MB/s]
Downloading:  88%|████████▊ | 459M/523M [00:05<00:00, 86.9MB/s]
Downloading:  89%|████████▉ | 468M/523M [00:05<00:00, 86.8MB/s]
Downloading:  91%|█████████ | 476M/523M [00:05<00:00, 86.8MB/s]
Downloading:  93%|█████████▎| 484M/523M [00:05<00:00, 86.9MB/s]
Downloading:  94%|█████████▍| 493M/523M [00:05<00:00, 87.2MB/s]
Downloading:  96%|█████████▌| 501M/523M [00:06<00:00, 83.9MB/s]
Downloading:  97%|█████████▋| 509M/523M [00:06<00:00, 84.9MB/s]
Downloading:  99%|█████████▉| 518M/523M [00:06<00:00, 85.6MB/s]
Downloading: 100%|██████████| 523M/523M [00:06<00:00, 87.1MB/s]
BaseModelOutputWithPastAndCrossAttentions(last_hidden_state=tensor([[[ 0.0530, -0.0137, -0.2393,  ..., -0.1245, -0.1116,  0.0225],
         [ 0.2470,  0.2260,  0.0397,  ...,  0.2413,  0.4349,  0.1768],
         [ 0.0436,  0.2554,  0.3014,  ..., -0.1825,  0.0148,  0.1905],
         [ 0.4704, -0.2377, -1.9606,  ..., -0.1460, -0.2216,  0.7026],
         [ 0.1739, -0.3471, -0.3398,  ...,  0.2080,  0.1483,  0.1136]]],
       grad_fn=<ViewBackward0>), past_key_values=((tensor([[[[-0.9706,  1.6740,  0.9128,  ..., -1.4558, -0.7133,  1.2801],
          [-1.8348,  2.4955,  1.7497,  ..., -1.5397, -2.3685,  2.4482],
          [-2.3145,  2.7101,  1.5073,  ..., -0.5781, -1.9292,  2.2634],
          [-2.0964,  2.4620,  2.3346,  ..., -0.5512, -0.7760,  2.0927],
          [-2.2980,  2.1683,  2.3144,  ..., -1.4571, -1.6000,  2.0501]],

         [[-0.0685,  0.6146, -1.0237,  ...,  0.1441,  1.8630,  0.3670],
          [ 0.6739, -1.3429, -0.1082,  ..., -3.4649,  3.4113,  0.9992],
          [-0.7140, -1.6105, -2.9748,  ..., -1.6920,  4.4462,  0.1892],
          [ 0.1616,  0.0461, -1.5248,  ..., -1.6373,  4.1694,  2.0673],
          [-0.0714, -1.5339, -0.2041,  ..., -0.6177,  2.3923,  0.1098]],

         [[-0.1442,  0.0673,  0.9683,  ..., -1.2523, -1.5758,  0.6272],
          [ 0.4345,  0.2073,  0.3242,  ..., -2.4565,  0.1195,  1.9386],
          [ 0.3226, -0.0288,  0.3313,  ..., -3.1311,  0.2430,  1.3300],
          [-0.2629, -0.4834,  0.2369,  ..., -3.3041,  0.3439,  2.5793],
          [ 1.0726, -0.1383,  1.0367,  ..., -1.8666,  1.3793,  1.9545]],

         ...,

         [[ 0.6158, -0.0193, -0.2532,  ...,  0.2937,  0.7897,  0.5587],
          [ 0.2512,  0.1103, -0.0282,  ...,  0.8245,  0.3106,  0.5934],
          [-0.0874, -0.1036, -0.0071,  ...,  1.2412,  0.6223,  0.4742],
          [-0.6714,  0.1988, -0.0682,  ...,  1.7253,  0.9976,  0.1185],
          [ 0.2586, -0.1179, -0.0624,  ...,  1.3856,  0.2490,  0.4159]],

         [[ 1.5653,  1.4870, -0.3752,  ..., -0.1429,  1.1948, -1.3825],
          [ 0.8506,  0.6267, -0.7537,  ..., -0.9063,  0.8328, -0.5789],
          [ 0.9876,  0.8250, -0.1679,  ..., -1.2784,  0.9794, -1.0050],
          [ 0.9305,  0.6660, -1.0610,  ..., -1.2928,  0.5044, -0.4814],
          [ 1.4465,  0.3536, -0.9735,  ..., -0.7088,  0.6830, -0.0933]],

         [[ 0.6844,  0.0054,  0.2316,  ..., -0.5452,  0.3212,  1.5770],
          [ 1.2434,  0.0677,  0.0617,  ...,  0.7098,  0.5259,  1.7975],
          [-0.1124,  1.0617,  0.0329,  ...,  0.2203,  0.6739,  1.2803],
          [-1.2424, -0.6885, -0.7169,  ...,  0.0725, -0.5476,  2.2931],
          [-0.4550, -0.1715,  0.5293,  ..., -0.0302,  0.8030,  1.0699]]]],
       grad_fn=<PermuteBackward0>), tensor([[[[ 1.7590e-01,  9.9564e-02,  1.0256e-01,  ..., -3.1547e-02,
            1.8605e-02, -1.3680e-02],
          [ 2.3196e-01, -2.3861e-01,  2.9457e-01,  ..., -5.8200e-02,
            7.4808e-02, -8.7832e-02],
          [-4.9511e-02,  7.4070e-03, -2.8154e-02,  ...,  1.2485e-01,
           -3.5409e-02, -7.5963e-02],
          [-2.2153e-01,  4.0724e-02, -2.4738e-01,  ..., -2.6879e-02,
            8.2440e-03,  1.3346e-01],
          [-9.9050e-02,  8.6200e-02, -1.4924e-01,  ...,  1.0040e-01,
           -1.3620e-01, -3.9094e-02]],

         [[ 4.7854e-01,  1.0785e-01, -2.5545e-01,  ..., -6.6785e-01,
           -2.6871e-01,  1.8379e-01],
          [ 5.5125e-01,  7.0201e-02,  1.0743e-01,  ...,  1.1353e-02,
            2.2987e-01, -7.8098e-02],
          [ 7.6954e-01, -1.2836e-01,  3.4602e-01,  ..., -1.1713e-01,
            3.3321e-01,  1.3056e-01],
          [ 4.2734e-01, -1.2921e-01, -9.4827e-02,  ...,  2.3222e-01,
           -4.5765e-02,  1.4689e-01],
          [ 6.8766e-01, -8.7459e-03,  1.6498e-01,  ..., -2.5414e-02,
            4.4872e-01, -7.0749e-02]],

         [[-4.6220e-02, -1.3808e-01,  3.5708e-02,  ...,  1.6425e-02,
            8.7861e-02, -2.4282e-02],
          [ 9.5054e-02,  9.0989e-02,  5.3470e-02,  ..., -2.7077e-02,
            9.1225e-02,  1.0601e-01],
          [-4.1603e-01, -3.7761e-02, -5.1928e-01,  ..., -1.2976e-01,
           -9.8224e-02,  2.0073e-01],
          [ 4.3101e-02,  2.1820e-02, -2.3061e-01,  ...,  5.1015e-02,
            1.4078e-02, -2.8447e-01],
          [ 1.3159e-02,  1.4886e-01,  5.2293e-02,  ...,  1.6578e-01,
            1.2342e-01,  1.4065e-01]],

         ...,

         [[-5.5756e-02, -1.0117e-01, -3.3120e-02,  ..., -1.2168e-01,
            1.8244e-02, -2.3754e-02],
          [-4.7730e-01,  5.3618e-01,  1.1470e-01,  ...,  2.1382e-01,
           -4.0779e-01,  1.0846e-01],
          [-1.4000e-01, -1.9747e-01, -1.0718e-01,  ...,  1.7734e-01,
           -2.0112e-01,  4.0236e-01],
          [-3.4627e-01,  4.2308e-01, -4.5653e-02,  ..., -1.2454e-01,
           -1.9140e-01, -4.5647e-04],
          [ 4.0457e-02, -2.2150e-01,  1.7805e-01,  ..., -8.0880e-02,
           -5.8263e-02, -8.3997e-02]],

         [[ 8.5613e-02, -1.5359e-01, -2.0006e-01,  ...,  1.5995e-01,
            1.4234e-01, -6.6704e-02],
          [-3.5043e-01,  7.0391e-02,  2.1141e-01,  ..., -6.4529e-01,
           -1.4333e-01,  8.3923e-02],
          [-2.1430e-01,  7.4153e-02, -1.3433e-01,  ...,  2.8439e-01,
            4.4380e-02, -1.0576e-01],
          [ 3.1361e-01,  1.3757e-01, -7.3767e-02,  ..., -6.3906e-01,
           -1.8905e-01,  2.1825e-01],
          [ 8.1954e-02, -2.4840e-02,  2.2433e-02,  ...,  2.9686e-02,
           -9.7372e-02,  5.5338e-02]],

         [[-8.6568e-02, -3.8672e-01,  9.7209e-02,  ...,  7.6971e-02,
           -2.6879e-01, -9.8475e-02],
          [-5.8485e-02, -1.2240e-02, -3.8542e-02,  ...,  1.7494e-01,
            3.7288e-01,  1.2658e-01],
          [-8.7157e-02, -1.6566e-02, -7.1012e-02,  ...,  1.6309e-01,
            3.8762e-02,  3.1650e-02],
          [ 2.2815e-01,  1.4950e-01, -2.8827e-01,  ...,  3.4470e-03,
           -1.3536e-01,  2.2376e-01],
          [ 1.1608e-01, -5.3995e-01, -5.1241e-02,  ..., -7.3520e-02,
            4.0055e-02, -6.9866e-02]]]], grad_fn=<PermuteBackward0>)), (tensor([[[[-1.5662e-01,  1.6423e+00, -1.3348e+00,  ...,  1.6332e+00,
           -1.3762e+00,  9.9192e-01],
          [ 8.6928e-01,  2.5685e+00, -1.6883e+00,  ..., -7.0216e-01,
           -2.2906e+00,  3.8237e-01],
          [ 8.1318e-01,  1.4673e+00, -1.3720e+00,  ..., -2.9314e-02,
           -1.5038e+00, -3.0176e-01],
          [ 2.9803e-01,  1.0934e-01,  1.5447e-01,  ..., -5.8906e-01,
           -1.6266e+00, -1.1208e+00],
          [ 7.1111e-01,  3.6426e-01, -7.4793e-01,  ..., -9.9811e-02,
           -1.2109e+00,  2.1198e-01]],

         [[-6.4042e-01, -2.4661e-01, -7.1877e-01,  ..., -2.5732e-01,
            6.5508e-01, -5.1007e-01],
          [-5.5587e-01,  6.0833e-01, -1.3343e+00,  ..., -6.4794e-01,
            8.7302e-01, -2.2356e-01],
          [-9.2801e-01,  5.3899e-01, -1.2114e+00,  ..., -1.0585e+00,
            4.2768e-01, -3.4744e-01],
          [-3.0372e-01, -2.2134e-01, -1.6153e+00,  ..., -7.0597e-01,
           -4.7869e-01, -1.2304e+00],
          [ 4.7420e-01,  2.0550e-02, -7.0636e-01,  ..., -6.5123e-01,
           -3.3610e-01, -3.4144e-01]],

         [[ 2.3560e-01,  1.0942e-01, -9.0574e-02,  ..., -1.3031e+00,
            8.6825e-02, -1.3109e-01],
          [-1.3051e-01,  1.6705e-01, -1.2011e-01,  ..., -1.0769e+00,
           -1.6215e-01,  4.1204e-01],
          [-4.9272e-02,  8.0602e-02, -1.2228e-01,  ..., -8.3993e-01,
            1.1068e-01,  2.5764e-01],
          [ 2.3408e-01,  3.9583e-01, -1.3784e-01,  ..., -1.0889e+00,
           -2.5668e-03,  1.9296e-01],
          [ 1.6254e-01,  5.0256e-02, -7.0564e-02,  ..., -1.3377e+00,
            2.1885e-02,  5.5437e-01]],

         ...,

         [[-1.8886e-01, -5.8206e-01, -7.7181e-01,  ..., -4.6246e-01,
            3.1854e-01, -4.2386e-01],
          [-1.7431e-01,  2.4567e+00,  1.4314e+00,  ..., -1.4023e-01,
           -2.7420e-01,  2.4984e-01],
          [-6.4468e-01,  6.5272e-01,  1.2134e+00,  ..., -9.2777e-01,
            1.3683e+00, -4.6400e-01],
          [-2.2817e+00,  8.7845e-01,  2.6406e+00,  ..., -1.9250e-01,
           -1.1961e+00,  4.4677e-01],
          [-2.0348e-01,  1.3825e+00,  2.0572e+00,  ...,  4.8078e-01,
           -1.7296e+00,  2.8359e-02]],

         [[-1.1482e+00, -2.8465e+00,  1.7656e-01,  ...,  1.7027e+00,
            1.6648e+00, -1.6975e+00],
          [ 1.7462e-01,  9.2828e-01, -4.5800e-01,  ..., -9.1707e-01,
            4.6858e-01, -4.4127e-01],
          [ 3.7161e-01,  4.0407e-01, -4.6503e-01,  ..., -3.8024e-01,
            4.6518e-01, -2.7998e-01],
          [-4.4134e-02,  7.6254e-01, -5.8167e-01,  ..., -8.0448e-01,
            5.5364e-01,  7.3189e-02],
          [ 2.2224e-01,  1.4123e-01, -5.4562e-01,  ..., -4.6865e-01,
            5.7062e-01, -1.5188e-01]],

         [[ 1.1212e+00,  2.1238e+00,  1.4072e-01,  ..., -5.4778e-01,
           -1.9993e-01,  7.4922e-01],
          [ 1.2248e+00,  1.8456e+00, -5.0574e-01,  ...,  1.3934e+00,
            1.3200e-01, -1.1804e+00],
          [ 4.7375e-01,  2.4056e+00,  7.4228e-01,  ...,  1.2234e+00,
           -1.3614e+00,  4.9688e-01],
          [-7.7144e-01,  2.6304e+00,  7.9746e-01,  ..., -8.2535e-01,
           -1.4945e+00, -1.0042e+00],
          [-1.5933e-01,  1.6211e+00,  2.2573e-01,  ..., -9.3011e-02,
           -6.5056e-01,  8.3301e-02]]]], grad_fn=<PermuteBackward0>), tensor([[[[ 1.5402e-01, -2.3608e-02, -3.3318e-02,  ...,  2.3801e-01,
            1.3243e-01,  2.4174e-01],
          [ 1.4298e-01,  1.8767e-01, -1.2800e-01,  ..., -2.2660e-01,
           -4.6244e-01, -3.8900e-02],
          [-1.4993e-01,  9.1094e-01, -3.7083e-01,  ...,  2.5402e-01,
           -5.0396e-01,  1.3749e-01],
          [-1.7675e-01, -1.6337e-01, -4.0210e-01,  ..., -1.9392e-01,
            7.0721e-02, -5.7761e-01],
          [ 7.9679e-01,  2.7786e-03, -6.1952e-02,  ...,  3.4274e-02,
           -1.4165e-01,  1.1635e-01]],

         [[ 3.1489e-01, -7.1358e-02,  6.2685e-02,  ...,  5.0654e-02,
           -6.5887e-01,  2.2787e-01],
          [ 1.7540e-01,  5.1056e-01,  7.9930e-01,  ..., -3.7897e-01,
            6.8204e-02,  3.5904e-01],
          [-9.6462e-02,  2.4428e-01,  4.9614e-01,  ...,  5.3062e-02,
            9.5801e-01,  4.2249e-01],
          [ 5.8537e-01, -5.0085e-01,  2.1125e-01,  ..., -5.7149e-01,
            1.7560e-01, -3.5075e-01],
          [ 3.2788e-01, -3.2853e-01, -2.4813e-01,  ...,  4.1474e-01,
            2.1426e-01,  2.5449e-01]],

         [[ 5.7193e-02, -2.4183e-01,  1.5918e-02,  ..., -5.8318e-01,
            1.6917e-01, -5.8768e-02],
          [ 4.9252e-01,  1.6655e-01,  4.7270e-01,  ..., -7.2064e-01,
            1.6249e-02,  5.4907e-02],
          [ 3.6085e-01,  4.2249e-02, -4.0496e-02,  ..., -7.4484e-01,
           -1.3859e-01,  1.0345e-01],
          [ 5.5564e-01, -5.0796e-02,  1.2230e-01,  ..., -6.6669e-01,
           -8.3693e-01, -2.0695e-02],
          [ 6.6719e-01, -1.0353e-01,  9.8830e-02,  ..., -4.8211e-01,
           -5.0656e-02, -5.6887e-01]],

         ...,

         [[-3.6945e-02,  5.9903e-01, -1.0183e-01,  ...,  4.2867e-02,
           -1.0561e+00, -2.2715e-01],
          [ 5.8844e-03, -1.1787e-01, -3.9837e-02,  ...,  2.0880e-02,
           -3.9521e-01, -3.6774e-01],
          [ 4.7640e-02,  2.2429e-01,  3.1230e-01,  ...,  2.3109e-01,
           -5.9350e-01, -2.6112e-01],
          [ 2.4943e-02, -9.7896e-01,  4.3425e-01,  ...,  3.4532e-01,
           -7.9044e-01, -3.8641e-01],
          [-3.0885e-01, -9.9148e-01,  1.5749e-01,  ...,  8.6608e-02,
           -7.8987e-01, -3.1548e-01]],

         [[ 3.9094e-01, -1.3596e-01, -2.0050e-01,  ...,  2.3203e-01,
           -3.5910e+00, -9.5989e-02],
          [-3.2107e-02, -1.1591e-01,  2.3247e-01,  ...,  4.6189e-01,
            1.3026e-01,  3.9747e-02],
          [-2.5550e-01, -2.0897e-01,  1.8833e-01,  ..., -2.1724e-01,
            4.1285e-02, -2.0056e-01],
          [ 8.0058e-02,  9.2737e-02, -1.9914e-01,  ...,  1.8804e-02,
           -1.2165e-01, -3.0129e-01],
          [-9.8699e-02,  9.6925e-02,  2.4426e-01,  ..., -1.9586e-01,
           -2.4658e-01,  1.7724e-01]],

         [[ 4.8408e-02, -1.8513e-01, -1.0398e-03,  ..., -2.2922e-01,
            1.9274e-01, -8.0920e-03],
          [ 4.6654e-02, -1.2114e-01, -1.5640e-01,  ..., -8.4000e-02,
            3.6298e-01,  1.9655e-01],
          [-1.5252e-02,  1.2396e-01, -2.1693e-01,  ..., -2.2036e-01,
            2.7967e-01,  3.0851e-01],
          [ 2.7916e-02,  4.1591e-01, -1.1142e-01,  ..., -2.7857e-01,
            2.1671e-01,  3.3429e-01],
          [ 1.2126e-01, -1.4811e-01,  1.9768e-01,  ..., -1.4386e-01,
            2.0777e-01,  3.3451e-01]]]], grad_fn=<PermuteBackward0>)), (tensor([[[[-0.1318, -1.0625,  0.2204,  ..., -0.6300, -0.1460, -0.0675],
          [ 1.3924, -1.6237, -1.6025,  ..., -0.5887,  0.2327, -2.4521],
          [ 0.4357, -1.5152, -0.9647,  ..., -0.1834, -0.1293, -1.1447],
          [-0.7615, -2.8303,  2.0162,  ..., -0.8280, -0.4874, -0.4754],
          [ 0.0358, -2.3034, -0.3780,  ...,  1.0244, -0.6885, -1.3657]],

         [[-0.5022,  0.3327, -0.4036,  ...,  1.2283, -0.5633, -0.4907],
          [-1.2102, -0.6699, -1.7375,  ...,  0.4006,  1.2817, -0.4390],
          [-1.8517, -0.7577, -0.1918,  ...,  0.0796,  0.8429,  1.3817],
          [-1.6611, -1.1005, -0.7044,  ...,  0.0491,  1.0727, -0.5113],
          [-1.2719, -1.5429, -1.6842,  ...,  0.6211,  0.9460, -1.1377]],

         [[ 1.3363,  3.0190,  3.7360,  ...,  0.6252,  1.7230, -0.7133],
          [-3.4435,  2.5479, -2.9284,  ..., -2.3357,  3.7772,  1.9482],
          [-4.0007,  0.2591, -2.6353,  ..., -3.6363,  4.1335,  0.4190],
          [-2.3891,  1.1513, -5.0157,  ..., -2.8085,  3.5907,  0.5094],
          [-3.7349, -0.2174, -4.2109,  ..., -3.2995,  3.2691,  0.8862]],

         ...,

         [[ 1.3163, -2.7678, -2.6654,  ...,  0.9293,  0.4377,  2.7041],
          [-1.7005,  2.4502,  0.4656,  ..., -0.7922, -2.4385,  0.3338],
          [-2.7072,  1.6356, -0.0909,  ..., -0.0870, -2.5314,  0.3139],
          [-2.3806,  1.8975,  0.9586,  ..., -0.8334, -2.1649, -0.6907],
          [-3.7194,  2.8237,  1.5085,  ..., -0.1043, -1.6027,  0.2228]],

         [[ 1.7277,  0.4507,  0.9105,  ...,  0.0304, -0.9858, -0.2994],
          [ 2.1653,  1.2410,  0.6264,  ...,  0.1824, -1.8031, -1.4786],
          [ 2.0792,  0.7471,  0.9823,  ...,  0.1123, -2.1957, -1.5933],
          [ 2.3667,  0.5518,  1.5805,  ..., -0.0361, -1.5019, -1.1254],
          [ 2.0166,  0.6256,  1.1297,  ...,  0.0694, -2.3993, -1.2250]],

         [[-0.2600,  0.1490, -0.5906,  ...,  0.2806,  0.2541,  0.2020],
          [-0.9068,  0.4040, -0.0709,  ...,  0.5576,  0.4885, -0.0146],
          [ 0.1920,  0.1637, -0.3991,  ...,  0.2519, -0.2749,  0.4086],
          [-0.5511, -0.1158,  0.0237,  ...,  0.9051,  0.4423, -0.2540],
          [-0.2628, -0.3767, -0.5705,  ...,  0.8204,  0.4021,  0.2640]]]],
       grad_fn=<PermuteBackward0>), tensor([[[[-0.0044, -0.0060, -0.1380,  ...,  0.0152,  0.0409, -0.5546],
          [ 0.4033, -0.2793, -0.3719,  ..., -1.1452, -0.1672,  0.6636],
          [ 0.6011,  0.7689, -0.0710,  ..., -0.2412,  0.0042,  0.7650],
          [ 0.6864, -0.8019,  0.4667,  ..., -0.6391,  0.2029,  0.6839],
          [-0.0060,  1.0654, -0.8982,  ..., -0.0063,  0.4145, -0.7409]],

         [[ 0.0196, -0.0310,  0.0566,  ..., -0.0130, -0.0177,  0.0153],
          [-0.2858,  0.1269,  0.0773,  ..., -0.2264,  0.2264, -0.2485],
          [-0.2367,  0.0108,  0.2280,  ..., -0.0505, -0.1735,  0.1840],
          [-0.5326, -0.1382, -0.6184,  ..., -0.8046, -1.2463,  0.1880],
          [-0.1685,  0.0110,  0.3741,  ..., -0.2904, -0.5199,  0.6103]],

         [[ 0.0060, -0.7748, -0.0544,  ...,  0.0816,  0.0211, -0.0381],
          [ 0.1551, -0.9633,  0.1427,  ...,  0.5207, -0.1885,  0.5032],
          [-0.0391, -1.4867,  0.1410,  ...,  0.1253, -0.0590,  0.3089],
          [ 0.3349, -1.7521,  0.0458,  ..., -0.6123, -0.1065,  0.3723],
          [ 0.1863, -1.4079, -0.1028,  ..., -0.1724,  0.0180,  0.0948]],

         ...,

         [[ 0.0713, -0.0434,  1.3028,  ..., -0.0181,  0.1826, -0.0156],
          [-0.0538, -0.7626,  2.2640,  ...,  0.3869,  0.3856,  0.1500],
          [ 0.2769,  0.2162,  2.5539,  ..., -0.0035,  0.2974, -0.1092],
          [-0.6548, -0.4305,  1.3475,  ...,  0.1106,  0.1324, -0.3254],
          [-0.3205, -0.6440,  2.8343,  ...,  0.0415, -0.2410,  0.3175]],

         [[ 0.0356, -0.1023, -0.1762,  ...,  0.1363,  0.1299,  0.1679],
          [ 0.7698,  0.2207, -0.3308,  ...,  0.1549, -0.7043, -0.7966],
          [ 0.4104,  0.1778, -0.3120,  ...,  0.3962,  0.0805, -0.2839],
          [ 0.1608, -0.4995, -0.0712,  ...,  0.8954, -0.0624, -0.4888],
          [ 0.9503,  0.3142,  0.7612,  ..., -1.1214, -0.2541, -0.3069]],

         [[ 0.0240,  0.0262,  0.0306,  ..., -0.0164,  0.2089,  0.0446],
          [-0.1774, -0.6164,  0.1278,  ..., -0.2834, -1.8231,  0.7639],
          [-0.1459, -0.6881, -0.4283,  ..., -0.0966, -1.9560,  0.1888],
          [-0.0632, -0.7851, -1.2836,  ...,  0.4466, -1.7303,  0.2600],
          [-0.0968, -0.1436, -0.6267,  ...,  0.5008, -1.2909,  0.0547]]]],
       grad_fn=<PermuteBackward0>)), (tensor([[[[ 4.8766e-02, -2.1665e-01,  1.5719e-01,  ..., -8.9878e-01,
            7.2553e-01, -1.1940e+00],
          [-1.0356e+00, -1.8130e-01, -1.3397e+00,  ..., -2.2399e-01,
           -8.9882e-01,  1.3276e+00],
          [-8.8831e-01,  4.0787e-01, -4.2665e-01,  ..., -9.1335e-01,
           -1.2772e+00,  1.0378e+00],
          [ 3.4834e-01, -8.0103e-01,  1.2843e+00,  ...,  5.0213e-01,
           -1.6890e+00, -4.7486e-01],
          [-8.4600e-01, -2.6589e+00,  3.2149e-01,  ...,  4.4007e-01,
            4.5740e-01, -4.3332e-01]],

         [[ 7.9398e-01,  2.0936e-01,  1.9167e-02,  ..., -1.6355e-01,
           -1.0906e+00, -1.8766e-01],
          [-1.0863e-01, -1.6188e+00,  2.2055e+00,  ...,  1.4346e+00,
            3.9374e+00,  2.0398e+00],
          [-5.3117e-01, -1.0060e+00,  7.3980e-02,  ...,  7.3000e-01,
            5.6576e+00,  9.4412e-01],
          [ 4.6430e-01, -9.4526e-01, -4.3511e-01,  ..., -1.1962e-01,
            5.5894e+00,  1.8238e+00],
          [-2.1476e+00, -1.3452e+00,  1.0562e+00,  ...,  3.6170e-01,
            4.8742e+00,  1.7200e+00]],

         [[ 3.4977e-01, -3.7168e-01, -3.2786e-01,  ...,  3.5171e-01,
            1.4341e+00,  2.5982e-01],
          [ 4.2102e-01, -5.8982e+00, -1.3955e+00,  ..., -3.6772e+00,
           -2.0719e+00, -5.4712e+00],
          [-4.1802e-02, -7.0493e+00, -1.4417e+00,  ..., -2.8737e+00,
           -3.7756e+00, -6.1265e+00],
          [-2.8095e-01, -5.0041e+00, -1.4907e-01,  ..., -4.0106e+00,
           -3.3702e+00, -5.1311e+00],
          [-7.2331e-01, -6.8279e+00, -1.4678e+00,  ..., -3.6394e+00,
           -1.9398e+00, -4.7794e+00]],

         ...,

         [[ 2.3129e-01,  1.7749e+00,  5.3036e-01,  ...,  2.6733e-01,
            4.8375e-01, -1.6757e+00],
          [ 2.0026e-01, -6.2583e+00,  2.1877e+00,  ..., -1.3897e+00,
           -1.6186e+00,  4.9409e+00],
          [-1.4066e+00, -5.9882e+00,  1.5971e+00,  ..., -1.5757e+00,
           -6.8382e-01,  5.7762e+00],
          [-1.5105e+00, -4.5623e+00,  2.3100e+00,  ...,  1.9863e-01,
           -7.6097e-01,  6.3563e+00],
          [ 3.9594e-01, -5.8773e+00,  1.3060e+00,  ..., -1.5648e+00,
           -1.1685e+00,  6.1503e+00]],

         [[ 4.6658e-02, -3.1419e-02,  1.2114e-01,  ..., -9.9283e-02,
           -1.0083e-01, -1.4549e-01],
          [-1.2447e+00, -1.6039e+00, -1.0188e+00,  ..., -1.7495e+00,
            8.2279e-01, -1.0526e+00],
          [-9.3251e-02, -2.7478e+00, -1.1833e+00,  ..., -1.6814e-01,
            1.8224e-01, -9.6570e-01],
          [ 2.7930e-01, -2.8319e-01, -4.4242e-01,  ..., -9.1226e-01,
           -6.2783e-01, -9.4037e-01],
          [-1.0931e-01, -1.9081e-01, -2.6465e-01,  ..., -2.2091e-01,
            3.4125e-01, -6.0161e-01]],

         [[ 3.9161e-01, -6.1812e-02,  1.9168e+00,  ..., -2.2250e-01,
           -2.1279e-01, -9.9491e-01],
          [ 3.0481e+00,  2.6143e+00, -2.2484e+00,  ...,  1.5371e+00,
            4.6030e-01,  2.3848e+00],
          [ 3.5652e+00,  1.4023e+00, -1.6737e+00,  ...,  8.3925e-01,
            1.4802e+00,  2.8034e+00],
          [ 2.0901e+00,  1.3682e+00, -1.9297e+00,  ...,  2.2705e+00,
           -3.8972e-01,  3.9399e+00],
          [ 3.8137e+00,  2.2652e+00, -3.7651e+00,  ...,  6.9402e-01,
           -3.7513e-03,  4.3839e+00]]]], grad_fn=<PermuteBackward0>), tensor([[[[ 4.9088e-02,  6.2656e-02, -6.1674e-03,  ...,  1.1223e-02,
            1.0092e-01,  2.7565e-02],
          [ 3.4844e-01, -1.0468e+00,  5.5674e-01,  ...,  8.4677e-01,
           -6.4409e-01, -5.2644e-01],
          [ 9.0298e-02, -3.9644e-01,  2.7643e-01,  ...,  4.5542e-01,
           -5.3742e-01, -6.9316e-01],
          [ 9.5429e-02, -6.4044e-01,  3.8654e-01,  ...,  8.8920e-01,
           -1.0769e+00, -1.5608e+00],
          [-4.7297e-01, -8.0739e-01,  4.6205e-01,  ...,  1.4032e-02,
           -6.4338e-02, -3.5396e-02]],

         [[-3.7407e-02, -6.4005e-03,  8.7001e-02,  ..., -4.7796e-02,
           -3.2202e-02, -4.0012e-02],
          [ 4.4432e-01,  2.3790e-01,  2.4416e-01,  ..., -1.4318e-02,
           -7.0716e-02,  1.4081e-01],
          [ 3.2779e-01,  5.8399e-02,  3.9499e-01,  ..., -2.8121e-01,
           -1.8764e-01,  2.4893e-02],
          [ 9.4433e-01, -6.5802e-02, -4.3887e-01,  ...,  4.9859e-01,
            7.8740e-01, -1.0636e-01],
          [-4.6391e-02,  7.1925e-03, -3.2034e-01,  ...,  8.3956e-02,
            5.0315e-01,  1.5096e-01]],

         [[ 3.9887e-02, -1.0315e-01, -4.1929e-02,  ..., -1.8554e-02,
            9.6552e-02, -1.5395e-01],
          [-5.8991e-01, -5.5089e-01, -9.4351e-01,  ...,  4.6885e-01,
            6.0745e-02, -5.5844e-02],
          [-6.4946e-02, -1.9911e-01,  1.1135e+00,  ...,  5.1064e-02,
           -1.8676e-01,  4.1609e-02],
          [ 4.7730e-01, -6.1616e-01, -6.0697e-02,  ..., -4.5071e-02,
           -6.5561e-01, -2.4879e-01],
          [ 6.6395e-01,  3.3562e-01, -9.9421e-01,  ...,  2.6696e-01,
            3.1598e-01, -4.5894e-01]],

         ...,

         [[-2.8601e-02,  1.2182e-01, -1.7708e-02,  ..., -2.6438e-02,
            5.9516e-02, -3.6307e-02],
          [ 3.1520e-01,  3.1634e-02, -9.8087e-03,  ..., -6.0115e-02,
           -2.8854e-01,  1.4823e-01],
          [-1.0675e-01,  1.3170e-01, -1.6675e-01,  ...,  1.3419e-01,
           -1.6130e-01,  7.2995e-02],
          [-4.7360e-01,  2.5233e-01,  2.4157e-02,  ...,  1.1482e+00,
           -1.1925e+00,  4.2617e-06],
          [-2.0872e-01,  4.7306e-01, -3.4658e-01,  ...,  4.0983e-01,
            1.8392e-02,  7.9382e-03]],

         [[-1.6626e-01, -1.4286e-01, -7.9067e-02,  ..., -2.5902e-01,
           -1.4665e-02, -3.0759e-02],
          [-1.0113e-02, -1.0856e+00,  7.5496e-01,  ...,  1.3461e-01,
           -8.8491e-02,  1.4906e+00],
          [ 2.7731e-01, -7.1459e-01,  7.0953e-01,  ...,  1.3236e-02,
            9.7033e-02,  1.0531e+00],
          [ 1.1717e+00, -1.2015e+00,  9.0899e-01,  ...,  5.7639e-02,
           -1.6985e-01,  7.3858e-02],
          [ 5.8210e-01,  1.0854e+00,  8.3712e-01,  ..., -2.1514e-01,
            4.3734e-01, -7.8206e-01]],

         [[ 1.1900e-01, -6.6088e-02, -4.2052e-02,  ..., -1.4421e-02,
           -8.6026e-02, -1.0313e-01],
          [ 9.6541e-02,  2.7553e-01, -3.2425e-01,  ..., -4.2882e-01,
           -2.2745e-01,  2.3476e-01],
          [ 2.2902e-01,  1.2116e-02, -1.8137e+00,  ..., -4.7918e-01,
           -1.8145e-01,  4.1571e-01],
          [-7.2667e-01,  2.5733e-01, -1.6341e-01,  ...,  6.9816e-01,
           -1.9473e-01,  2.5240e-01],
          [ 7.2513e-01,  8.6979e-02, -1.5850e-01,  ...,  8.4285e-02,
           -3.2990e-01,  2.6306e-01]]]], grad_fn=<PermuteBackward0>)), (tensor([[[[-8.8494e-01, -1.1705e-01,  3.3638e-01,  ..., -9.6571e-01,
            1.6086e-02, -2.9542e+00],
          [ 8.7922e-01,  1.6237e+00, -2.7220e+00,  ..., -1.9895e+00,
           -2.2475e+00,  6.5433e+00],
          [ 1.1498e+00,  1.2402e+00, -2.3427e+00,  ..., -2.2912e+00,
           -1.7495e+00,  6.6168e+00],
          [ 2.0055e+00,  1.5162e+00, -2.6148e+00,  ..., -1.4749e+00,
            9.0602e-01,  5.6495e+00],
          [-2.2821e-01, -1.0497e-01, -2.5007e+00,  ..., -2.0487e+00,
           -1.6907e+00,  8.2848e+00]],

         [[ 3.8577e-01, -6.8490e-02,  4.6245e-01,  ..., -1.4116e-01,
           -6.8129e-02, -2.2122e+00],
          [-1.6953e+00,  5.0828e-01,  2.6259e+00,  ..., -8.3102e-01,
           -7.5975e-01,  6.6576e+00],
          [-3.8885e+00, -9.8093e-01,  3.6157e+00,  ..., -1.2371e+00,
            7.5860e-03,  6.6705e+00],
          [-2.4971e+00, -7.7720e-01,  3.3081e+00,  ..., -8.8270e-01,
            8.5871e-02,  4.7817e+00],
          [-2.3747e+00,  1.8189e-01,  2.7463e+00,  ...,  1.7065e+00,
           -1.6088e+00,  6.1548e+00]],

         [[ 1.3703e-01, -6.5138e-01, -2.1594e-01,  ...,  1.4911e-01,
            2.6310e-01, -1.7072e-01],
          [ 4.6264e-01,  1.8531e+00,  1.6080e+00,  ...,  8.3421e-01,
            1.2447e+00,  1.2077e-01],
          [-1.5158e+00,  2.3226e+00,  1.6826e+00,  ..., -3.4110e-01,
            9.1427e-02,  1.0164e+00],
          [-1.6825e+00, -4.2120e-01,  2.9897e-01,  ..., -5.1527e-01,
            6.4858e-02, -6.4533e-01],
          [-2.5971e-01,  2.6257e+00,  6.8382e-01,  ...,  1.2026e-02,
           -1.2671e+00, -1.4449e+00]],

         ...,

         [[-3.9712e-01,  1.3165e-02,  1.2875e-03,  ...,  1.2579e+00,
            6.0933e-02,  1.7770e+00],
          [ 1.3343e+00, -7.2030e-01, -5.1513e-01,  ..., -2.5676e+00,
           -1.1343e+00, -1.2415e+00],
          [ 1.6986e+00, -1.1409e+00, -5.1925e-02,  ..., -3.4870e+00,
            4.5648e-01, -9.0337e-01],
          [ 1.2578e+00, -4.3136e-01, -2.0769e+00,  ..., -3.3061e+00,
           -8.2535e-01, -1.6187e+00],
          [-2.3508e-01, -9.2000e-01, -1.8668e+00,  ..., -3.8580e+00,
           -1.1501e+00, -1.0528e-01]],

         [[-3.4194e-01, -1.2436e-01,  2.1257e-01,  ...,  2.5841e-01,
           -3.5488e-02,  2.4488e-02],
          [ 8.3284e-01,  2.7459e-01,  9.5044e-01,  ...,  8.0016e-01,
            6.1839e-01, -1.5041e+00],
          [ 6.6688e-02,  3.1421e-01,  3.2119e-01,  ...,  1.0004e+00,
           -4.9638e-01, -3.3743e-01],
          [-2.5250e-01,  5.3944e-01,  1.4781e+00,  ...,  2.5685e-01,
            2.7114e-01,  3.6381e-01],
          [ 2.2998e-01, -5.1351e-01,  1.2536e+00,  ...,  6.4712e-01,
           -6.7503e-01, -2.3498e-01]],

         [[ 3.4159e+00,  2.1540e+00, -2.1205e+00,  ..., -2.8551e+00,
           -3.8837e+00, -1.1908e+00],
          [-5.1380e+00, -1.5247e+00,  2.4941e+00,  ..., -1.6254e+00,
            1.1478e+01,  1.6499e+00],
          [-7.8050e+00, -2.4687e+00,  1.2786e+00,  ..., -3.1693e+00,
            1.2727e+01,  2.1972e+00],
          [-6.2529e+00, -4.0362e-01,  2.8779e+00,  ..., -2.0485e+00,
            6.3220e+00, -1.2608e+00],
          [-6.0076e+00, -1.8160e+00,  6.1665e+00,  ..., -2.3327e+00,
            9.7334e+00, -3.0119e+00]]]], grad_fn=<PermuteBackward0>), tensor([[[[-0.0096, -0.0519,  0.0195,  ...,  0.0595,  0.0299,  0.0704],
          [ 0.9772,  0.2666, -0.3315,  ..., -1.2055,  0.5357, -0.4633],
          [ 0.2826, -0.0410,  0.0277,  ..., -0.1198,  0.2637,  0.0713],
          [-1.6780,  0.1257,  0.0214,  ..., -0.5101, -0.1718, -0.7016],
          [-0.5040,  0.3606,  0.1849,  ...,  0.1181,  0.0361,  0.0527]],

         [[-0.0681, -0.0157, -0.1464,  ..., -0.0466,  0.0438, -0.0242],
          [ 0.8736,  0.2970, -0.0071,  ..., -0.0495, -0.1670,  0.2172],
          [ 0.0157,  0.0040,  0.2625,  ...,  0.0482, -0.1384,  0.3871],
          [ 0.0533, -0.0248, -0.3261,  ...,  0.7589, -0.1830,  0.7068],
          [-0.0711, -0.1694,  0.1085,  ..., -0.5241,  0.2944, -0.4246]],

         [[ 0.0645,  0.1003,  0.0918,  ...,  0.0161, -0.0875, -0.0067],
          [-0.2457,  0.3270, -1.0554,  ...,  0.7323, -0.2182,  0.2848],
          [-0.7481,  0.4380, -0.6971,  ...,  0.5318, -0.3274, -0.7197],
          [-0.2736,  0.8033, -0.2764,  ..., -0.3137,  0.4019,  0.1637],
          [ 0.3955,  0.7612, -0.0470,  ...,  0.2236, -0.1068,  0.4775]],

         ...,

         [[-0.0083,  0.0804, -0.0829,  ...,  0.0491,  0.0347, -0.1400],
          [ 0.6573,  0.5186, -0.2037,  ..., -0.1756,  0.3145,  0.7512],
          [-0.0768,  0.1822,  0.5083,  ..., -0.0965,  0.0815,  0.2934],
          [-1.1065,  0.3834, -0.5550,  ..., -0.7877, -0.3019,  0.6682],
          [-0.1850,  0.6478, -0.6281,  ...,  0.0381,  0.0447, -0.2501]],

         [[-0.1300, -0.0529,  0.1044,  ..., -0.0675,  0.0469, -0.0103],
          [-0.6747, -0.6985, -0.2033,  ...,  0.0168,  0.2096,  0.1116],
          [-0.2319, -0.1853, -0.0180,  ..., -0.2475,  0.6775,  0.0269],
          [ 0.5872, -0.6089, -1.5718,  ...,  0.2216,  0.1597,  0.6948],
          [ 1.0053, -0.4880, -0.1548,  ..., -0.3616,  0.1325, -0.0835]],

         [[-0.0137, -0.0078, -0.0289,  ..., -0.0246,  0.0099, -0.0152],
          [-0.6180,  0.0845, -0.2192,  ..., -0.1918,  0.2803, -0.0235],
          [-0.1593,  0.2884, -0.6303,  ..., -0.4634,  0.4622,  0.3242],
          [-0.0025,  0.2431, -0.1563,  ..., -0.7715,  0.2807,  0.2144],
          [ 0.7167,  0.3752, -0.0947,  ..., -0.5741, -0.0444,  0.6885]]]],
       grad_fn=<PermuteBackward0>)), (tensor([[[[ 3.2784e-02, -3.0164e-01,  2.2821e-01,  ...,  1.6954e+00,
           -2.1624e-01, -7.5642e-02],
          [ 1.6108e+00,  9.2464e-01,  7.4591e-01,  ..., -4.2026e+00,
           -2.2189e-01, -1.4860e+00],
          [ 7.3925e-01,  8.9980e-01,  1.2081e+00,  ..., -3.7170e+00,
           -2.6006e-01, -1.9648e+00],
          [-3.0593e-02,  7.8754e-01,  9.1603e-01,  ..., -3.7605e+00,
            1.0331e-01, -1.0918e+00],
          [-1.7863e+00,  6.0353e-01, -1.6878e-01,  ..., -5.1717e+00,
           -2.1356e-01,  3.8194e-01]],

         [[ 1.6445e-01,  9.7323e-01, -1.4239e+00,  ..., -1.2736e-01,
            2.6054e-01,  9.2129e-01],
          [-2.1263e+00, -5.8295e+00,  2.2762e+00,  ..., -2.4167e+00,
           -4.3525e-01, -3.4747e+00],
          [ 1.1715e+00, -5.4696e+00,  1.8708e+00,  ...,  1.7774e-01,
           -1.3082e+00, -2.6058e+00],
          [-2.7758e-01, -4.6695e+00,  1.3966e+00,  ..., -8.2469e-01,
           -1.3299e+00, -1.5987e+00],
          [ 4.8986e-01, -5.8022e+00,  4.3565e+00,  ..., -1.9137e+00,
            8.1013e-01, -2.6240e+00]],

         [[-6.6771e-01,  2.4735e-01, -4.6270e-02,  ...,  1.7921e-01,
            4.6848e-02, -2.8902e-01],
          [ 1.0005e+00,  7.6726e-02, -6.3028e-01,  ..., -1.4184e+00,
           -1.4159e+00,  1.3787e-01],
          [ 3.0189e+00,  1.2265e+00,  1.3839e-01,  ..., -1.2720e+00,
           -1.1207e+00, -6.0942e-01],
          [ 3.1529e+00, -9.1017e-01, -7.5620e-01,  ..., -8.2353e-01,
           -1.1433e-01,  6.2328e-01],
          [ 9.0134e-01,  4.1126e-01, -9.6940e-01,  ..., -1.4677e+00,
            2.3810e-01,  1.9366e+00]],

         ...,

         [[-4.0053e-02,  1.2722e-01,  1.4467e-01,  ..., -9.4189e-02,
            2.2058e-02,  1.5769e-01],
          [-6.3529e-01, -1.3811e-01, -8.8769e-01,  ...,  1.5349e+00,
            1.9831e-01, -4.7695e-02],
          [ 5.2656e-01,  5.9109e-01,  2.8637e-01,  ...,  2.0854e+00,
            2.5085e-01, -1.8108e-01],
          [ 1.8517e+00, -1.8718e-01,  1.3077e-01,  ...,  1.9755e+00,
           -1.0787e+00,  6.5873e-01],
          [ 1.3111e+00, -4.3745e-01, -9.2446e-01,  ...,  1.7096e+00,
           -3.6738e-01,  4.7958e-01]],

         [[-3.0033e+00,  3.9400e-01, -2.4926e-02,  ..., -4.8616e-01,
           -3.3859e-01,  1.2360e+00],
          [ 5.4850e+00, -5.6572e-01, -1.4467e+00,  ..., -1.4157e-01,
            1.7902e+00, -1.3817e+00],
          [ 4.8411e+00,  2.2152e-01, -1.0374e+00,  ..., -3.4201e-01,
            4.3440e-01, -9.6590e-01],
          [ 5.0392e+00, -4.6671e-01, -6.9487e-01,  ..., -7.3891e-01,
           -6.7736e-01, -1.9242e+00],
          [ 5.1982e+00, -9.2724e-01, -1.7033e+00,  ..., -1.2516e+00,
            8.1261e-01, -1.5033e+00]],

         [[-5.6354e-03, -2.4766e-01, -2.5135e-03,  ..., -1.6960e-01,
            3.2356e-01,  7.7852e-02],
          [-3.2433e-02, -2.5221e+00,  9.0981e-01,  ..., -1.3292e+00,
            5.6196e-02, -1.2525e+00],
          [-3.4994e-01, -2.4945e+00,  1.7282e-01,  ..., -5.1078e-01,
            7.9858e-01, -1.2843e+00],
          [-1.3952e+00, -4.0069e-01,  1.0538e+00,  ...,  3.2851e-01,
           -5.6107e-01, -1.0815e+00],
          [-2.1934e-01, -2.3399e+00,  1.0217e+00,  ...,  1.9279e-01,
            6.0801e-01, -5.4505e-01]]]], grad_fn=<PermuteBackward0>), tensor([[[[-2.3297e-02, -2.2363e-02,  4.0103e-03,  ...,  2.3276e-03,
           -3.5343e-02,  3.4658e-01],
          [ 1.7878e+00, -2.0226e-01,  5.3187e-01,  ..., -5.1709e-02,
            1.4288e-01, -9.2569e-01],
          [ 8.9776e-01, -2.2820e-01,  7.0814e-01,  ...,  6.4914e-02,
            7.8736e-01, -9.0007e-01],
          [ 4.5384e-01,  7.0735e-01,  2.2922e-01,  ...,  1.0098e-01,
            7.5938e-01, -1.1460e+00],
          [ 6.1267e-02,  4.3366e-01, -1.3490e+00,  ...,  8.0903e-01,
           -4.6388e-01, -1.9364e-01]],

         [[ 4.7413e-03, -2.2730e-02,  1.2142e-02,  ..., -1.7133e-02,
            2.4038e-02,  1.1147e-02],
          [ 6.2605e-01, -1.2681e-01,  4.6525e-01,  ..., -9.1562e-01,
            8.1387e-01, -3.3404e-01],
          [ 1.0077e+00, -1.8241e-01, -6.5852e-01,  ..., -2.5416e-01,
            1.5129e+00,  5.8085e-02],
          [ 1.0615e+00, -9.3878e-01,  6.9884e-01,  ..., -2.7794e-01,
            1.7145e+00, -4.3421e-02],
          [ 4.1729e-01,  1.0709e-01, -4.4126e-01,  ..., -1.4493e+00,
            1.6497e+00,  2.2936e-01]],

         [[-5.9155e-02,  2.9002e-03, -3.8955e-02,  ..., -4.3641e-02,
            1.3672e-02, -8.2469e-02],
          [-7.5135e-04, -2.5521e-01, -7.4059e-01,  ..., -1.4025e+00,
            5.4869e-01,  1.1923e-01],
          [-3.9955e-01,  4.1394e-01, -1.7664e-01,  ..., -8.4038e-01,
            7.9501e-01,  2.5979e-01],
          [-7.2326e-01,  2.3251e-01, -1.2385e-01,  ..., -2.3025e-01,
           -3.5558e-01,  5.7998e-01],
          [-6.9278e-01,  1.7771e+00,  9.5392e-01,  ..., -9.5790e-02,
           -7.2291e-01, -1.7692e-01]],

         ...,

         [[-3.1947e-01, -1.8975e-01, -6.3943e-02,  ..., -4.6983e-01,
            2.2903e-01,  1.0098e-01],
          [ 1.8197e+00, -1.8336e+00,  1.3286e+00,  ...,  1.9690e+00,
           -2.9223e-01,  1.7737e-01],
          [ 1.5096e+00, -9.1225e-01,  1.2883e-01,  ...,  9.3922e-01,
           -7.6452e-01,  6.0677e-01],
          [ 2.6054e+00, -4.1233e-01,  1.1730e-01,  ...,  2.3199e-02,
           -2.5344e+00,  1.1474e+00],
          [ 1.9458e+00, -3.9425e-01, -1.6890e-02,  ..., -8.0653e-02,
           -1.0912e-03, -1.3616e+00]],

         [[-7.8436e-02, -1.3601e-01, -4.0731e-02,  ..., -1.8876e-01,
           -1.4006e-01,  1.2306e-01],
          [-1.3453e-01, -6.4662e-01,  6.6342e-01,  ..., -4.5152e-01,
           -5.8993e-01, -8.4304e-01],
          [-2.6158e-01, -3.4778e-01, -8.2513e-02,  ..., -9.6671e-01,
           -5.3280e-01, -9.2661e-02],
          [-5.5134e-01, -7.1199e-02,  2.4435e-01,  ..., -6.9977e-01,
           -1.5694e-01, -1.0693e+00],
          [-3.1745e-01, -2.6903e-01, -3.8605e-01,  ..., -7.2463e-02,
           -5.3896e-01, -2.2516e-01]],

         [[-3.1061e-02, -3.6021e-02,  9.0767e-02,  ...,  7.4807e-02,
           -3.6683e-02,  1.8671e-02],
          [-1.2921e-01, -4.1537e-01, -3.3070e-01,  ..., -4.0818e-01,
           -6.0640e-01,  3.1346e-01],
          [-4.4102e-02, -9.6212e-02, -8.5648e-01,  ..., -2.7890e-02,
            2.8478e-01,  4.5391e-01],
          [-2.6616e-01,  1.2863e-01,  9.6675e-01,  ..., -4.4183e-01,
            1.3013e+00, -5.1109e-01],
          [-2.0684e-01,  3.6566e-01, -1.0952e+00,  ...,  5.3827e-01,
           -8.8518e-02,  2.3647e-01]]]], grad_fn=<PermuteBackward0>)), (tensor([[[[-3.2637e-01,  8.7869e-01, -1.7338e-01,  ...,  1.1323e+00,
           -1.7099e-01,  1.3227e-01],
          [ 4.0682e-01, -4.2113e+00, -1.4068e+00,  ..., -3.6183e+00,
           -4.4289e-01,  8.6498e-01],
          [ 4.4289e-01, -4.5109e+00, -1.9845e+00,  ..., -3.3165e+00,
            5.1283e-01,  3.2089e-01],
          [ 8.4120e-01, -3.4927e+00, -6.5901e-01,  ..., -3.4589e+00,
            1.0680e+00,  1.1760e+00],
          [ 1.8809e-01, -4.5319e+00,  3.4496e-03,  ..., -4.3399e+00,
            2.1186e+00,  7.4519e-02]],

         [[ 5.9638e-02,  8.5713e-01, -6.3980e-01,  ..., -2.9613e-02,
            2.9474e-01,  1.2961e-02],
          [ 7.7426e-01, -3.4914e-01,  1.6956e+00,  ...,  1.8312e+00,
            2.0048e-01, -1.2981e+00],
          [-2.7189e-01,  1.1024e-02,  1.0201e+00,  ...,  2.4932e+00,
            3.7676e-02, -3.3382e-01],
          [ 9.4470e-01,  5.4816e-01,  1.4242e+00,  ...,  1.0948e+00,
           -1.3178e+00,  6.8701e-01],
          [-3.1536e-01, -2.2370e-01, -2.9986e-01,  ...,  1.7112e+00,
           -3.8295e-01, -6.2328e-01]],

         [[-3.1808e-01,  1.1994e-01, -9.8032e-01,  ..., -3.5461e-01,
           -6.8214e-02, -1.3295e-01],
          [-4.4920e-01, -1.3175e-01,  3.2521e+00,  ...,  2.9216e-01,
           -6.3339e-01, -2.9502e-01],
          [-1.0203e+00,  4.9793e-01,  2.9459e+00,  ...,  2.7080e-01,
           -1.0528e+00,  9.2279e-01],
          [-5.5592e-01,  2.4681e-01,  2.6699e+00,  ...,  6.5269e-01,
            1.9334e-01,  9.8206e-01],
          [-3.0186e-02, -1.9546e-01,  1.9469e+00,  ...,  5.9395e-01,
            4.4278e-01, -8.6679e-01]],

         ...,

         [[ 3.8336e-01,  8.2223e-02, -7.3157e-02,  ..., -5.1192e-02,
            2.2547e-01,  1.4341e-02],
          [-2.8259e-01, -1.9201e+00, -1.1183e+00,  ...,  3.5498e-01,
            1.0319e+00, -6.4936e-01],
          [ 3.2330e-01, -2.9246e+00, -1.1234e+00,  ..., -4.0797e-01,
            1.0676e+00, -4.1374e-01],
          [-5.2200e-01, -7.1518e-01, -1.0027e+00,  ..., -9.7230e-01,
            1.5876e+00, -1.9422e-01],
          [ 6.4574e-03, -1.3672e+00, -1.9331e+00,  ..., -1.0124e+00,
           -5.6214e-02,  4.3803e-01]],

         [[ 2.0240e-01,  6.5908e-02,  3.2197e-01,  ...,  4.1540e-01,
            1.4468e-02,  2.3231e-01],
          [ 1.2523e+00,  5.9905e-01,  8.8246e-02,  ..., -8.6822e-01,
           -5.0354e-01,  4.3219e-01],
          [ 1.0033e+00, -1.4375e-01,  5.9073e-01,  ..., -5.4477e-01,
            8.4845e-02, -1.7197e-01],
          [ 2.4207e+00, -3.0575e-01,  7.5101e-01,  ..., -7.9357e-01,
           -9.7652e-01,  4.0906e-01],
          [ 1.9084e+00, -7.7618e-01,  1.6903e+00,  ..., -8.3242e-01,
           -1.8942e+00,  1.1117e+00]],

         [[-3.0075e+00,  5.4783e-01,  5.5633e-01,  ..., -9.3269e-01,
            3.1288e-01,  2.0191e-01],
          [ 8.1386e+00,  6.7674e-01, -2.2141e+00,  ...,  1.6099e+00,
            3.8929e-01, -1.8135e-01],
          [ 9.1399e+00,  5.5120e-01, -2.2615e+00,  ...,  1.9101e+00,
           -2.5173e-01,  5.3780e-01],
          [ 7.3972e+00,  7.1368e-01, -2.5716e+00,  ...,  1.4241e+00,
           -1.6825e+00, -8.9060e-01],
          [ 8.8053e+00,  4.2708e-01, -1.5595e+00,  ...,  2.7597e+00,
           -1.3309e+00, -1.4891e+00]]]], grad_fn=<PermuteBackward0>), tensor([[[[ 3.6105e-02, -5.0920e-02,  1.9079e-02,  ..., -7.0462e-02,
            7.8206e-04, -9.5287e-02],
          [-8.2560e-01, -1.6433e-01,  3.2615e-01,  ...,  1.5180e-02,
            4.8105e-01, -3.1854e-01],
          [-1.7452e-01,  1.2412e-02, -2.2870e-01,  ...,  4.4079e-01,
            4.5812e-01, -1.2639e+00],
          [ 3.6470e-01,  5.1354e-01,  5.9135e-01,  ...,  1.1807e-01,
            3.0826e-01, -7.4766e-01],
          [-5.7955e-02,  4.0697e-01,  6.1814e-02,  ..., -5.6123e-03,
           -2.7953e-02, -1.0808e+00]],

         [[ 6.7154e-02,  2.7802e-02, -3.1921e-02,  ..., -3.4743e-02,
            2.0332e-02, -5.9948e-03],
          [ 6.4384e-01,  7.3428e-01, -6.4294e-01,  ..., -5.6763e-02,
            5.1062e-01,  2.0320e-01],
          [ 2.0241e-01,  6.1742e-01, -2.6660e-01,  ...,  3.0636e-01,
           -2.3723e-01,  3.9112e-01],
          [ 3.4536e-01, -1.4546e-02,  5.4485e-01,  ...,  9.0210e-01,
           -4.3854e-01,  1.2136e-01],
          [-4.6642e-01,  1.7622e+00, -1.7134e-01,  ...,  4.6734e-01,
           -9.6902e-01, -3.4786e-01]],

         [[ 7.5267e-02,  1.3920e-02,  3.6354e-03,  ...,  2.8545e-02,
           -7.1940e-02, -6.7323e-02],
          [-2.3199e-01,  5.0063e-02, -2.1456e-01,  ..., -8.5352e-01,
            4.9608e-01, -2.4874e-01],
          [-3.5004e-01, -3.7469e-01, -7.8334e-01,  ..., -8.0984e-01,
            1.1186e+00, -1.2292e-02],
          [ 1.0247e+00, -6.1152e-01,  6.0260e-01,  ...,  3.1542e-01,
            3.9953e-01, -7.1116e-02],
          [ 5.9923e-01, -3.7654e-01,  3.0572e-01,  ..., -1.4295e-01,
            1.0610e+00,  2.1563e-01]],

         ...,

         [[ 1.7294e-03,  2.5408e-02,  2.0194e-02,  ..., -6.8257e-02,
           -2.6157e-02,  2.0736e-02],
          [-5.5511e-01,  8.1093e-02, -2.6873e-01,  ...,  1.3403e+00,
           -2.8910e-01, -1.3519e-01],
          [-5.9936e-01,  3.8204e-01, -2.8838e-01,  ...,  8.7584e-01,
           -1.8204e+00, -4.7055e-01],
          [-6.6890e-01, -5.4211e-01, -7.6068e-01,  ...,  1.1513e+00,
           -5.4273e-01,  2.2256e-01],
          [-9.8153e-01, -2.1624e+00, -1.6029e-01,  ...,  5.5969e-02,
            7.6084e-01, -2.0222e-02]],

         [[ 4.2696e-02, -7.5619e-03,  1.9380e-02,  ...,  2.8514e-02,
           -1.2640e-02,  9.3482e-04],
          [ 9.1049e-01,  1.4645e+00, -8.0264e-01,  ...,  6.6169e-01,
            6.0504e-01,  7.5308e-01],
          [ 1.1664e+00,  4.4888e-01,  4.2179e-01,  ...,  5.8827e-01,
            2.9115e-01, -4.4443e-01],
          [-7.6207e-01,  9.4271e-02, -1.6601e-01,  ...,  5.9498e-01,
            8.5246e-01, -1.0566e-01],
          [-6.0164e-02, -3.1022e-01,  6.8694e-01,  ..., -7.1013e-01,
           -6.2186e-01,  4.5495e-01]],

         [[ 6.9054e-02, -1.9165e-01, -7.4112e-02,  ..., -3.6272e-02,
            1.9359e-01, -4.5177e-02],
          [-4.2709e-01, -3.2601e-01, -1.0170e+00,  ...,  9.7749e-01,
           -5.0469e-01,  6.8743e-01],
          [-1.3850e-01,  2.9038e-03, -7.8623e-01,  ...,  7.0431e-01,
           -6.7867e-01,  1.0110e+00],
          [ 1.6974e-01, -1.6448e-01, -2.7440e-01,  ...,  4.4319e-01,
            1.0121e+00,  3.5337e-01],
          [-4.0246e-01, -7.2136e-01, -2.2601e-01,  ..., -4.7839e-01,
            2.7870e-01,  4.6428e-01]]]], grad_fn=<PermuteBackward0>)), (tensor([[[[ 1.0376e+00, -2.4648e-01, -1.5264e-01,  ...,  6.3285e-01,
            7.2599e-01, -3.1369e-01],
          [-4.7676e+00, -3.2713e+00, -4.6545e-02,  ..., -4.9860e-01,
           -5.0155e+00,  2.7614e-01],
          [-4.3598e+00, -2.8035e+00, -1.4492e+00,  ...,  2.5620e-02,
           -4.2017e+00, -1.9942e-01],
          [-3.5033e+00, -5.9694e-01, -8.3350e-01,  ..., -1.4505e+00,
           -3.9481e+00,  3.8880e-01],
          [-3.8167e+00, -1.5106e+00, -1.0508e+00,  ..., -1.2307e+00,
           -2.9675e+00,  2.2575e-01]],

         [[-1.4668e-01, -6.6439e-02,  1.5897e-01,  ..., -4.3018e-02,
           -8.7371e-01, -1.9589e-01],
          [-1.1093e+00,  6.6262e-01, -1.0539e+00,  ..., -5.5684e-01,
           -1.0412e+00,  3.8443e-01],
          [-1.3144e+00,  1.0730e+00, -6.2434e-01,  ..., -5.7737e-01,
           -1.0939e-01,  3.9126e-01],
          [-1.7404e+00,  1.3938e+00, -9.7743e-01,  ..., -1.7429e+00,
           -1.2020e+00, -1.9696e-01],
          [-1.0631e+00,  9.1892e-01, -9.8092e-01,  ..., -2.0835e+00,
           -1.5144e+00, -7.6347e-01]],

         [[ 1.8450e-01,  3.1127e-01,  1.1326e+00,  ..., -4.6201e-01,
            4.3714e-01, -5.0350e-01],
          [-2.0086e+00, -1.1766e+00, -1.0570e+00,  ..., -1.1925e+00,
           -1.9625e+00,  1.0180e+00],
          [-1.9097e+00, -1.6713e+00, -2.0410e+00,  ..., -4.7677e-01,
           -2.8514e+00,  3.2336e+00],
          [-8.0545e-01,  2.1527e-01, -5.9493e-01,  ..., -3.6546e-01,
           -2.0042e+00,  2.8835e+00],
          [ 1.2326e-01,  2.5759e+00, -2.9319e+00,  ...,  3.8286e+00,
           -3.4094e+00,  8.0914e-01]],

         ...,

         [[-1.5956e-01,  6.2100e-02, -2.3516e-01,  ...,  2.3788e-03,
            1.6357e-01,  1.0968e-02],
          [-2.5161e+00, -1.4926e-01, -9.9057e-01,  ...,  6.3516e-01,
            7.3840e-01, -8.4298e-01],
          [-1.6555e+00, -9.0446e-01, -8.1934e-01,  ...,  1.7268e+00,
            1.1985e+00, -6.1387e-01],
          [-1.6482e+00, -1.8944e+00, -9.2101e-01,  ..., -1.0306e-01,
            3.2358e-01, -5.0422e-01],
          [-1.6691e+00,  3.6330e-01, -7.8001e-01,  ..., -1.1134e+00,
           -1.9934e-01, -3.5791e-01]],

         [[-3.4752e-01, -2.1773e+00,  1.1447e-01,  ..., -8.6873e-02,
           -4.3887e-02,  9.1306e-01],
          [ 1.1407e-01,  2.6037e+00, -1.5496e+00,  ..., -1.7522e+00,
           -3.5838e-02, -3.4268e-01],
          [-4.1974e-01,  3.9355e+00, -1.8980e+00,  ..., -1.6527e+00,
           -2.6531e+00,  1.6956e+00],
          [ 5.1421e-01,  2.8418e+00, -8.8639e-01,  ..., -3.7614e-01,
           -2.4758e+00,  9.6414e-01],
          [ 4.6669e-01,  3.6112e+00, -1.1484e-01,  ...,  2.5912e+00,
           -1.6105e+00, -9.5305e-01]],

         [[ 3.7213e-01,  7.0688e-02, -1.4030e-01,  ...,  6.5445e-01,
            1.3776e-01,  2.5245e-01],
          [-7.5425e-01, -1.1270e+00, -5.6842e-02,  ..., -1.1020e+00,
            4.8513e-01, -1.5070e+00],
          [-3.5186e-01, -4.1867e-01,  5.1747e-01,  ..., -5.4417e-01,
           -9.7728e-01, -1.1426e+00],
          [-1.2946e+00, -4.9512e-02,  1.0634e+00,  ..., -1.1202e+00,
            9.9057e-01, -9.2959e-01],
          [-7.8605e-01,  3.6546e-01,  4.9948e-01,  ..., -7.3540e-01,
            6.3933e-01, -4.6611e-01]]]], grad_fn=<PermuteBackward0>), tensor([[[[-3.0836e-02,  4.1770e-02, -5.5116e-02,  ..., -1.2211e-02,
            8.0840e-04,  1.9999e-02],
          [ 1.1057e+00,  1.6903e-01, -9.5124e-01,  ...,  6.6509e-01,
            4.5961e-01, -2.3908e-01],
          [ 5.9558e-01,  3.5296e-01,  4.8080e-01,  ...,  3.9347e-01,
            2.6027e-01, -9.1538e-01],
          [ 3.0211e-01, -4.3492e-01,  1.7188e-01,  ...,  4.2687e-02,
           -1.1858e-01,  5.1285e-01],
          [-1.3767e-01,  2.0870e-01, -6.7562e-01,  ..., -7.1594e-01,
           -1.3690e-01,  4.9661e-01]],

         [[ 8.5170e-03, -2.5506e-02,  3.0265e-02,  ...,  1.9872e-02,
           -4.4784e-02,  1.8725e-02],
          [-2.9388e-01,  9.8014e-02, -9.8786e-01,  ...,  3.9735e-01,
            2.1016e-01,  2.8024e-01],
          [-4.0205e-01,  1.4434e-01, -3.5802e-01,  ...,  5.2803e-01,
            8.3174e-01,  1.1678e+00],
          [ 1.1445e+00, -1.0995e+00, -7.5140e-01,  ..., -4.3948e-01,
           -7.6491e-01,  1.1398e-01],
          [ 6.4464e-01, -5.8727e-01, -1.0397e+00,  ...,  4.0126e-02,
           -3.5195e-01,  4.5275e-01]],

         [[ 4.3662e-02, -1.8857e-02,  5.6767e-02,  ...,  2.0677e-02,
           -2.2482e-04,  7.5198e-03],
          [-2.1630e-01,  9.9355e-01, -3.2740e-01,  ..., -7.9229e-01,
           -6.1090e-02, -2.8375e-01],
          [-7.1417e-01, -4.3306e-01,  1.5530e-01,  ...,  1.1651e+00,
            5.4687e-01,  2.1804e+00],
          [-5.0647e-01, -1.5844e-01, -1.5427e+00,  ...,  5.9105e-01,
            2.8160e-01, -7.8062e-01],
          [-1.8383e+00, -8.3039e-02, -1.1768e+00,  ...,  8.4435e-01,
            2.3457e-01,  4.6268e-01]],

         ...,

         [[-1.9144e-01,  8.7269e-02,  5.6517e-02,  ...,  4.2666e-02,
            3.9368e-02, -1.3311e-01],
          [-6.1943e-02,  4.6646e-01, -2.8178e-01,  ..., -4.4757e-02,
            8.6745e-01, -8.8913e-02],
          [-2.6362e-01,  8.6503e-01, -1.9244e-02,  ..., -1.0001e-01,
            5.5903e-01, -7.1435e-01],
          [ 3.2403e-01, -5.2341e-01, -1.2559e+00,  ...,  1.2480e+00,
            3.0835e-01, -2.2864e+00],
          [ 2.3782e-01, -6.7701e-01, -6.9363e-02,  ..., -5.9599e-01,
            4.3591e-01, -2.3052e-01]],

         [[-5.8905e-01, -8.0886e-03,  4.7519e-02,  ..., -9.9036e-03,
            1.5850e-02, -6.0772e-03],
          [-1.0845e+00,  3.5836e-01, -5.0078e-01,  ...,  2.3958e-01,
            5.7615e-01,  7.0684e-01],
          [-1.2403e+00,  2.6270e-01, -9.3001e-01,  ...,  9.9400e-01,
            3.1610e-01,  8.9827e-01],
          [-1.6026e+00, -4.9238e-01, -1.9647e-01,  ...,  8.3614e-01,
            1.2156e+00, -3.5790e-01],
          [-1.5941e+00,  3.2820e-01, -4.7823e-01,  ...,  1.4124e-01,
           -1.7781e-02,  9.1007e-01]],

         [[ 5.2297e-03,  7.8464e-02, -3.9536e-02,  ...,  5.4487e-02,
            4.0338e-02, -4.4751e-02],
          [-4.6119e-01,  8.5600e-01, -8.6741e-01,  ..., -2.9191e-01,
           -1.3538e+00, -9.6075e-02],
          [-6.3030e-01,  1.4630e-01, -5.9291e-01,  ...,  2.1067e-01,
           -3.2658e-01, -1.2420e+00],
          [-1.4685e-01,  4.5310e-01, -6.6578e-01,  ..., -4.2283e-01,
           -2.2941e+00, -1.3154e+00],
          [ 8.4889e-02, -8.6649e-02, -1.0079e+00,  ..., -2.7713e-01,
            7.0672e-01,  6.5931e-02]]]], grad_fn=<PermuteBackward0>)), (tensor([[[[-0.0354, -2.3398,  0.1632,  ..., -0.2238, -0.1893,  0.0734],
          [-1.0761,  4.9946,  0.4478,  ..., -0.2685, -0.4752, -0.7377],
          [-0.0808,  3.4150,  1.0626,  ..., -0.6483,  0.6249, -0.5982],
          [-0.8450,  4.9785, -0.9404,  ..., -0.3035, -0.3512, -0.1920],
          [-0.3097,  4.4646, -0.1116,  ..., -0.4222,  0.6816, -1.0945]],

         [[-0.7973,  0.2199,  0.4667,  ..., -0.5184,  1.0617,  1.1078],
          [ 0.4751,  0.2889, -0.0321,  ...,  0.3508,  0.3673, -1.3043],
          [ 0.9751,  0.5485,  0.8113,  ...,  0.4703, -0.2315, -1.4501],
          [ 0.7917,  0.4870,  0.6286,  ...,  0.2404,  0.3601, -0.5501],
          [ 0.0732,  0.6929,  0.9934,  ...,  1.3629,  1.4581, -0.3628]],

         [[-0.8498,  0.4698,  0.0261,  ...,  0.4947, -0.2308,  1.1487],
          [ 1.0158, -0.7828,  1.0488,  ...,  0.0159,  0.7942,  0.2791],
          [ 0.8138, -0.6384,  0.3970,  ...,  1.2297,  0.9968,  0.2116],
          [ 1.6742, -0.5450,  0.5703,  ...,  1.1663,  0.6349, -0.3270],
          [ 0.3878, -1.2577,  1.3819,  ...,  0.6651,  1.3689, -0.5181]],

         ...,

         [[-0.2984, -0.1281,  0.1453,  ...,  0.1904,  1.7339, -2.8597],
          [ 1.5347,  0.1231, -0.5401,  ..., -0.3106, -4.8969,  5.6338],
          [ 1.5007,  0.8672, -0.5233,  ..., -0.4986, -4.8510,  5.4111],
          [-0.4454, -0.1532, -0.8568,  ..., -0.0793, -4.7887,  4.7712],
          [ 0.7245,  0.8172, -0.5348,  ..., -0.6393, -5.3460,  5.5635]],

         [[ 0.1910,  0.3602,  0.2196,  ..., -0.2211,  0.0312, -0.1513],
          [-0.8380, -0.5265,  0.0073,  ...,  0.0539, -0.0160, -0.2554],
          [-1.0735, -0.5641, -0.0898,  ..., -0.1863,  0.2185, -0.1890],
          [-0.4489, -1.1593, -0.3829,  ...,  1.9596,  0.7516,  0.1214],
          [ 0.4726, -0.6231,  0.3862,  ...,  2.1031,  0.1787,  0.6572]],

         [[ 0.3625,  0.1039,  0.6137,  ...,  0.5196,  0.5744, -0.3359],
          [ 0.2853, -1.6101, -1.3272,  ..., -1.6917, -3.8798, -0.0182],
          [ 0.2868, -0.9251, -1.1646,  ..., -1.8721, -3.6443, -0.5594],
          [ 0.9404, -2.0547, -0.2527,  ..., -1.9803, -2.6577, -0.4327],
          [ 0.1150, -1.4610,  0.0548,  ..., -1.5205, -3.3002,  0.1087]]]],
       grad_fn=<PermuteBackward0>), tensor([[[[ 6.5363e-02, -7.8252e-03, -1.6065e-02,  ...,  1.2438e-01,
           -6.8575e-02, -4.3273e-02],
          [ 2.4643e-02,  1.3026e-01, -1.6459e-01,  ...,  1.2624e+00,
            8.7694e-01,  2.8388e-01],
          [ 5.5928e-01, -7.9160e-01, -1.9387e+00,  ...,  1.1458e+00,
            1.4961e-01, -8.9907e-01],
          [-1.5965e+00,  1.9013e-01, -2.3367e-01,  ..., -5.5747e-01,
            4.4411e-01, -6.9862e-01],
          [-5.4846e-01,  1.1141e+00, -6.8398e-01,  ..., -4.5728e-01,
            9.7019e-01, -8.9291e-02]],

         [[ 8.1593e-03,  3.7725e-02,  5.9576e-02,  ..., -2.0037e-03,
           -7.2108e-04,  1.2033e-03],
          [-5.8590e-01,  3.7442e-01,  1.0020e+00,  ..., -1.2509e+00,
           -5.2379e-01, -1.1079e+00],
          [-6.4543e-01,  1.3011e+00,  1.2044e+00,  ...,  1.0402e+00,
            1.2569e-01,  4.8326e-01],
          [ 6.0140e-01,  2.1851e+00, -1.8104e-01,  ...,  6.9321e-01,
           -3.7005e-01, -1.2225e+00],
          [-1.5327e+00, -1.9626e-01, -1.6820e+00,  ..., -6.4909e-01,
            2.6130e-01, -1.5801e-01]],

         [[ 5.2740e-02, -4.0095e-02,  6.0241e-02,  ...,  5.5740e-02,
           -6.2179e-02, -6.5946e-02],
          [-1.8639e-01,  1.1358e-01,  3.7150e-01,  ...,  2.5474e-01,
            5.1546e-02,  3.9120e-01],
          [ 5.9791e-02, -7.7903e-04,  1.2699e-01,  ...,  3.4055e-01,
            1.8779e-02,  3.5182e-01],
          [ 3.3224e-01,  5.1684e-01,  1.1111e+00,  ...,  1.1556e+00,
            9.5961e-01, -3.0655e-01],
          [-3.1475e-01,  4.6871e-01,  1.1127e-01,  ..., -1.3752e-01,
           -6.5294e-02,  6.8532e-01]],

         ...,

         [[-8.8191e-02, -4.2089e-02,  3.7978e-02,  ..., -8.9379e-02,
            3.4805e-02,  1.4007e-02],
          [ 8.7396e-01,  3.5090e-01, -5.0685e-01,  ..., -5.5516e-01,
            5.3380e-01,  5.6518e-02],
          [ 4.6496e-01,  5.6087e-01, -2.7614e-01,  ..., -1.4456e+00,
            1.4470e-01,  6.1488e-01],
          [-3.4311e-01,  1.0670e+00, -4.5802e-01,  ..., -5.5069e-01,
            2.0408e+00,  3.0035e-01],
          [ 4.9590e-01, -8.1093e-02,  2.6610e-02,  ...,  4.4404e-02,
            1.8599e+00,  3.0370e-01]],

         [[ 1.3777e-01, -6.6464e-02,  1.3369e-01,  ...,  6.4657e-02,
            3.1961e-02, -1.2535e-01],
          [-1.6845e+00,  1.4819e-01, -3.0695e-01,  ...,  1.1595e+00,
            4.3650e-01,  1.4513e-01],
          [-1.6848e+00,  6.7543e-01, -3.7684e-01,  ...,  1.5262e-01,
            5.2445e-01,  9.5677e-02],
          [-1.8220e+00, -9.1123e-01, -1.8414e+00,  ..., -4.2590e-01,
            1.5639e-01,  1.2168e-01],
          [-1.5278e+00,  4.7719e-01, -7.3938e-01,  ..., -1.9454e-01,
            7.7875e-01,  7.3514e-02]],

         [[ 2.0971e-01, -5.6491e-02, -5.9956e-02,  ...,  3.7434e-02,
            3.9753e-02,  2.7075e-02],
          [-2.5491e-01, -4.0122e-01, -6.1966e-01,  ...,  2.6864e-01,
            2.5014e-01,  8.9296e-01],
          [-1.8362e-01, -3.1695e-01, -8.8638e-02,  ...,  1.3048e-01,
           -7.8617e-01,  8.1072e-01],
          [-2.0093e+00, -5.4969e-01, -3.3942e-01,  ...,  5.1620e-01,
           -3.7176e-01, -7.2491e-01],
          [-1.4515e+00,  4.8476e-01,  6.2712e-01,  ..., -7.3589e-02,
           -3.7878e-01,  1.7636e-01]]]], grad_fn=<PermuteBackward0>)), (tensor([[[[ 0.0272, -0.2537, -0.4484,  ...,  0.3188,  0.3232,  0.3653],
          [-0.6271, -0.4423, -0.9100,  ...,  1.9808, -0.4710,  0.0414],
          [-0.4736, -0.8674, -0.2990,  ..., -0.3213, -0.8216, -0.6593],
          [ 0.7599,  0.1851, -0.0829,  ...,  0.6689, -0.1617,  0.3159],
          [-0.3258, -0.1632, -0.4298,  ...,  0.5044,  0.3958,  0.8392]],

         [[-0.2788,  0.1551,  0.1127,  ...,  0.0435, -1.1459, -0.1482],
          [ 0.1111, -0.3842, -0.3989,  ..., -0.3827, -0.1900,  0.1569],
          [ 0.5694, -0.5989, -0.3237,  ...,  1.0557, -0.1809,  0.4039],
          [-0.8325,  0.2121,  0.8149,  ..., -0.7450, -0.6405, -0.2680],
          [-1.2164,  0.2591, -0.0897,  ...,  0.5341,  0.5261,  0.3878]],

         [[-1.2380, -0.1126,  0.5452,  ..., -0.6537,  0.4603, -0.2895],
          [ 1.2720, -0.3546, -0.6730,  ...,  1.4354, -0.2492, -1.0218],
          [ 1.5896, -0.1622, -1.1232,  ...,  0.8653, -1.0884, -0.1940],
          [ 1.1730,  2.2100,  0.5004,  ...,  0.9975, -0.5147,  0.5569],
          [ 0.8398,  1.5637, -0.0968,  ...,  0.2396, -0.7644,  0.3537]],

         ...,

         [[ 0.7999, -0.9007, -0.3972,  ..., -1.0478, -0.4211,  0.4906],
          [ 1.3921, -0.7203, -1.4222,  ...,  0.6655,  1.2048, -0.3536],
          [ 1.4705, -0.3559, -1.6317,  ...,  0.3660, -0.4593, -0.4413],
          [ 0.2389, -0.6677, -1.1275,  ...,  1.0864, -0.9589, -1.3112],
          [ 1.2800, -0.6959,  0.0185,  ...,  0.7997, -0.6251,  0.0855]],

         [[-0.9050,  2.5628,  0.3109,  ...,  0.3530,  1.9446, -0.5305],
          [-1.2991, -4.0310,  0.6955,  ...,  0.3445, -5.3756,  1.8677],
          [ 0.0755, -2.4890,  1.2219,  ..., -0.5825, -4.3110,  0.1312],
          [-0.4697, -2.6668,  1.3540,  ..., -0.0809, -5.6818,  0.9068],
          [-0.1801, -4.2980,  1.8520,  ..., -0.2059, -4.2477,  2.3378]],

         [[-2.0185, -0.3657, -1.1158,  ..., -0.3912,  0.0567,  0.2527],
          [ 2.7369,  0.1396,  2.0207,  ...,  1.3567, -0.4944,  1.2061],
          [ 2.4151, -0.5353,  0.2838,  ...,  0.3568, -0.0126,  0.6582],
          [ 1.8294,  0.7969,  1.6802,  ...,  0.3412, -0.7427,  1.1150],
          [ 1.7365,  0.6256,  2.2205,  ...,  0.3748, -0.9411, -0.1127]]]],
       grad_fn=<PermuteBackward0>), tensor([[[[-5.2829e-02, -8.9784e-02,  1.3019e-02,  ...,  1.1785e-01,
           -1.3861e-02,  2.9725e-02],
          [-3.8549e-01, -8.0779e-01,  9.4546e-01,  ...,  6.5094e-01,
            4.1517e-01,  1.0727e-01],
          [-6.3594e-02,  4.3115e-02,  7.3490e-02,  ..., -1.6117e-01,
           -1.0211e-01,  1.4340e-01],
          [-1.4665e+00, -1.1967e+00, -6.7182e-01,  ..., -3.0123e-01,
            1.3670e+00, -1.7387e+00],
          [-1.5258e+00, -3.1563e-01, -1.4074e-01,  ..., -1.5674e-01,
           -7.1805e-03, -2.2522e-01]],

         [[ 1.8538e-02,  1.3586e-02, -6.0780e-02,  ...,  2.7396e-02,
            8.9818e-03,  2.9315e-02],
          [ 8.8571e-02, -1.0145e+00, -2.4855e-01,  ..., -4.8304e-01,
           -4.4495e-01, -9.3407e-01],
          [ 6.7974e-01,  2.9080e-01,  7.6243e-01,  ..., -5.6191e-01,
           -2.3151e-01,  1.2456e-01],
          [ 2.6921e-01,  4.7441e-01,  9.6789e-01,  ...,  9.4126e-01,
            1.5023e+00, -1.7956e+00],
          [ 1.3102e+00, -1.0181e-01,  2.8465e-01,  ..., -4.0300e-02,
            9.7835e-01, -1.0044e+00]],

         [[ 4.3295e-02,  2.9719e-02, -7.1790e-02,  ...,  1.0435e-03,
           -1.6481e-02, -6.2361e-03],
          [ 4.9795e-01, -1.7375e-01,  2.2942e-01,  ...,  1.0204e+00,
           -6.7490e-01, -6.4816e-03],
          [ 7.0578e-01, -2.0810e-01,  3.6799e-01,  ...,  7.5344e-01,
            1.5151e-01, -3.3273e-01],
          [-4.4946e-01, -9.0646e-01,  6.5912e-01,  ...,  7.1667e-01,
            7.1878e-01,  9.1848e-01],
          [-5.5957e-01, -1.2276e-01,  3.5039e-01,  ...,  1.9431e-01,
           -4.2784e-01,  5.3941e-02]],

         ...,

         [[-1.8254e-02,  2.9206e-02, -1.6896e-02,  ..., -1.3794e-02,
           -2.8781e-02,  2.8856e-02],
          [ 5.8796e-01, -2.5001e-01,  2.1126e-02,  ...,  5.2231e-03,
           -9.5165e-01,  5.3361e-01],
          [-1.0081e+00,  8.3290e-01,  8.1058e-01,  ...,  6.5036e-01,
           -4.8494e-01,  1.7019e+00],
          [-3.9442e-01,  2.6294e-01,  9.1182e-01,  ..., -2.3218e+00,
           -8.8488e-01, -6.6869e-01],
          [-6.6947e-01, -1.0271e-02,  5.1051e-01,  ..., -5.6418e-01,
           -1.2361e-01, -5.9249e-01]],

         [[-6.9019e-02, -3.3031e-02,  2.6570e-02,  ...,  1.2109e-02,
           -4.8550e-02, -1.0738e-01],
          [ 1.1253e+00, -2.3586e-01, -4.0789e-02,  ...,  1.2250e+00,
            1.3039e-01, -3.4633e-01],
          [ 1.2573e+00, -4.9249e-01,  1.9398e-02,  ...,  9.7694e-01,
            5.9966e-01, -3.9594e-01],
          [ 8.4073e-01, -5.2249e-01, -6.7681e-01,  ...,  1.2013e+00,
           -1.3864e+00, -2.5545e-02],
          [ 2.4236e-01,  1.2010e-01,  2.8589e-01,  ...,  7.7146e-01,
           -7.6955e-01,  1.5447e+00]],

         [[-1.4664e-03,  4.0603e-02, -4.8897e-02,  ..., -1.3849e-02,
            1.6516e-02, -2.4615e-03],
          [ 1.8577e-01, -6.5145e-01, -6.5146e-01,  ...,  8.8917e-01,
            2.8380e-01, -9.4657e-01],
          [ 1.9158e-01,  9.6865e-02, -1.5682e-01,  ...,  6.9871e-01,
           -1.5953e-01, -7.5410e-01],
          [-3.3769e-01,  3.9950e-01,  9.3834e-01,  ...,  7.9275e-01,
            7.3601e-01, -7.4401e-01],
          [-7.2728e-01, -1.3377e+00, -1.1666e+00,  ...,  5.0475e-01,
           -2.6387e-02, -5.2085e-01]]]], grad_fn=<PermuteBackward0>)), (tensor([[[[-0.5069,  0.4825, -0.8580,  ..., -1.0225, -1.3198,  0.2109],
          [ 0.3025,  0.6047,  0.1297,  ...,  2.0929,  0.2885, -2.0715],
          [ 0.9374,  0.5086,  0.7627,  ...,  2.4422,  0.6873, -1.7763],
          [-0.1876,  0.2970,  0.3633,  ...,  0.6674,  0.6973, -1.8348],
          [ 0.8828,  1.3149,  1.2335,  ...,  1.4942,  0.0060, -0.7146]],

         [[ 0.8611, -2.0596,  0.1490,  ...,  0.2371, -2.4527, -0.4371],
          [ 1.2852,  2.0747, -0.6886,  ..., -0.1657,  1.4251, -0.0795],
          [ 0.9519,  1.2935, -1.2819,  ..., -0.2411,  2.0321,  0.4034],
          [ 1.6719,  1.9951, -1.2201,  ...,  1.3384,  1.6690, -0.7929],
          [ 1.3321,  1.7295, -0.5557,  ...,  0.0247,  1.9838, -0.5768]],

         [[ 1.0103,  0.3728, -0.1706,  ..., -0.8065, -1.3946, -0.3728],
          [-0.5095, -0.3942, -0.6458,  ...,  0.6060, -0.3780, -0.3256],
          [-0.1554, -0.9991, -0.2617,  ...,  0.7160, -0.3941, -0.8668],
          [-1.2057, -0.0957,  0.4046,  ...,  1.3611, -0.6942,  0.1711],
          [ 0.1775, -0.8493, -0.5805,  ...,  0.9317, -0.3098, -1.1783]],

         ...,

         [[ 0.2318, -0.5798,  0.4860,  ..., -0.6775,  1.0579,  0.2846],
          [-0.1493,  0.5973, -0.7958,  ..., -2.6057, -1.1053,  0.8619],
          [ 0.2289,  1.0747, -0.9906,  ..., -3.3273, -0.6746,  0.2586],
          [-0.6913,  0.6801, -1.0802,  ..., -1.9457, -2.6984,  1.0896],
          [-0.6008,  1.3586, -1.0831,  ..., -1.7779, -0.5403,  0.3634]],

         [[ 0.2475,  0.5588,  0.5510,  ...,  0.7237,  0.0790,  0.8426],
          [ 0.4245,  3.0943,  0.1308,  ...,  0.8383, -0.9669,  0.3766],
          [-0.5022,  3.5487,  0.3371,  ..., -1.0983, -1.1552,  0.0310],
          [-1.2201,  2.1956, -0.7848,  ...,  0.6124, -1.1433,  2.5727],
          [ 1.0573,  2.1385,  1.4337,  ..., -0.5333, -1.1279,  0.5484]],

         [[-0.7419,  0.3051, -1.6006,  ..., -0.4013,  0.2329, -1.3062],
          [-1.4629, -0.2797, -1.9916,  ..., -1.2279,  0.9010, -1.1898],
          [-1.5916,  1.0883, -1.5881,  ..., -0.4685,  1.8982, -1.1973],
          [ 0.4077, -0.3870, -1.3748,  ..., -0.8245,  0.6468,  0.3719],
          [ 0.7073, -0.4831,  0.6523,  ...,  1.9505,  2.9627,  0.2937]]]],
       grad_fn=<PermuteBackward0>), tensor([[[[ 8.3213e-03,  4.8852e-02, -8.0794e-02,  ...,  6.6222e-02,
           -3.9223e-02, -8.3081e-02],
          [-2.3462e-01, -2.1689e-01,  1.0137e+00,  ...,  5.8081e-02,
           -7.5950e-02, -3.8082e-01],
          [-4.9420e-01, -2.9674e-01, -3.8088e-01,  ...,  7.3919e-01,
            1.2360e-01, -1.4323e+00],
          [-6.7838e-01, -1.2847e+00,  1.5788e+00,  ...,  2.0256e+00,
           -3.0706e-02,  7.0765e-01],
          [ 2.7779e-01, -4.4231e-02,  9.9544e-01,  ..., -1.3998e-01,
           -5.7540e-01, -5.7264e-01]],

         [[ 4.9996e-02, -7.3457e-04,  5.3660e-02,  ..., -3.2779e-02,
           -3.1049e-02, -1.2545e-02],
          [ 8.7422e-01, -7.6256e-02, -1.1535e-01,  ..., -6.2978e-01,
            2.9268e-01,  4.7039e-01],
          [ 3.1951e-01, -1.1324e-01,  7.7048e-01,  ...,  6.6294e-01,
            1.1786e-01,  2.0289e-01],
          [ 2.7052e-01, -1.1987e-01,  9.4267e-01,  ..., -1.2170e+00,
            2.3035e+00,  1.4124e+00],
          [ 4.7093e-01,  9.7040e-01, -6.2696e-01,  ..., -6.1159e-01,
           -4.4966e-01,  1.5132e+00]],

         [[-3.9760e-03,  1.8928e-02, -2.3610e-02,  ...,  6.0530e-04,
            1.4498e-02,  6.2650e-02],
          [ 8.0977e-01,  1.4486e-01,  5.5658e-02,  ...,  8.7285e-01,
            1.0636e-01, -1.0256e+00],
          [-1.0156e-01,  9.3937e-01, -3.7614e-01,  ...,  2.9506e-01,
           -6.0209e-01,  3.9851e-01],
          [-1.5783e+00, -9.8152e-01,  9.3926e-01,  ..., -1.6962e-01,
            1.2794e+00, -7.6056e-01],
          [ 8.1572e-01, -7.9100e-01,  1.4132e+00,  ..., -7.8284e-01,
            2.7810e-01, -3.8184e-02]],

         ...,

         [[-9.1555e-02,  1.9912e-02, -9.6007e-03,  ..., -3.8988e-02,
           -1.2533e-02,  6.2787e-03],
          [-3.9613e-01,  8.1027e-01, -6.6337e-02,  ...,  1.7382e-01,
           -4.6225e-01, -8.0468e-02],
          [ 6.8796e-01, -1.3956e-01,  3.7697e-01,  ...,  5.5507e-01,
           -3.4882e-01, -2.3023e-01],
          [-1.1481e-01,  4.1130e-01,  1.4725e-01,  ...,  8.7707e-01,
           -5.8721e-01, -8.3445e-01],
          [-6.6966e-01, -9.0118e-01,  1.6282e-01,  ..., -8.9014e-01,
           -8.8027e-01,  4.9121e-01]],

         [[ 6.5674e-02,  1.4786e-02,  6.8458e-02,  ...,  2.4556e-02,
            5.6556e-02,  2.0550e-02],
          [-2.2145e-01,  9.0658e-01,  1.8033e+00,  ..., -2.1279e+00,
            2.0283e+00, -1.7435e+00],
          [-3.3140e-01, -1.7858e-01,  1.5818e+00,  ..., -9.5400e-01,
            3.3448e+00, -2.0647e+00],
          [ 1.1303e+00, -1.3449e+00, -3.0305e-01,  ...,  8.9020e-02,
           -4.8015e-01,  1.2203e-01],
          [-1.0663e-02,  7.9229e-01,  2.0276e+00,  ...,  7.8246e-01,
            5.5617e-01, -2.0390e+00]],

         [[-1.1217e-01,  2.3648e-02, -5.0429e-02,  ..., -8.5543e-02,
            7.1119e-02, -1.1717e-02],
          [ 2.0649e-01,  1.1595e-01, -1.3196e-01,  ..., -4.2364e-02,
            1.1823e+00,  6.2216e-03],
          [ 8.0884e-04, -1.4813e-01, -5.3331e-01,  ..., -1.9672e-01,
            9.0681e-01, -4.0601e-01],
          [-1.0777e-01, -3.8276e-01, -9.3460e-01,  ...,  4.4666e-01,
            1.5121e+00,  3.1829e-01],
          [-4.5711e-01, -2.0546e-01, -7.6122e-01,  ..., -1.4951e-01,
            3.3287e-01,  6.9401e-01]]]], grad_fn=<PermuteBackward0>)), (tensor([[[[-1.7161, -0.3134, -0.2944,  ...,  0.1691,  0.3235, -0.5194],
          [-0.4214, -0.4050, -0.4763,  ...,  0.8270, -0.7408,  0.0973],
          [-0.5198, -0.3677, -0.0095,  ...,  1.1533, -0.7225, -0.7734],
          [ 0.5599, -0.2347, -1.1768,  ...,  0.4925, -0.4893,  1.0560],
          [-0.4487, -0.1774, -0.9955,  ...,  0.1688, -1.7715, -0.1549]],

         [[ 0.1036, -0.0696,  2.2923,  ...,  0.2424,  0.0902, -0.1867],
          [ 0.9062, -0.3726, -0.3097,  ..., -0.0043,  0.8178, -0.0434],
          [ 0.1174, -0.8003, -0.9808,  ...,  0.7453, -0.1963, -0.9001],
          [-0.2972, -1.4146, -0.3341,  ...,  0.4544,  0.4238,  0.1093],
          [ 0.4281, -0.2350, -0.8483,  ...,  0.8430,  0.9144, -0.5829]],

         [[-0.1944,  1.0640,  0.4731,  ..., -0.5328,  0.3241, -0.1038],
          [-0.6372,  0.1499, -0.4102,  ...,  1.7041,  0.4894,  0.0193],
          [-0.3877,  0.4143,  0.1512,  ...,  1.2959,  0.9774, -0.3048],
          [-0.2410,  1.0406,  0.5100,  ...,  0.7326,  0.8054, -0.1469],
          [-0.4700,  1.0114, -0.6867,  ...,  0.5154,  0.3383, -0.1972]],

         ...,

         [[ 0.5581,  0.9737, -0.8686,  ..., -0.7255,  0.7067,  0.8475],
          [ 0.3736,  0.2205, -1.4620,  ..., -0.6006,  0.0569,  0.2796],
          [-0.2479,  0.5485, -0.0746,  ..., -1.0234, -0.2244, -0.6247],
          [ 1.6113,  0.9696, -0.7635,  ..., -0.7161,  0.7673, -0.8774],
          [ 0.5931, -0.1715, -0.7995,  ..., -0.3478, -0.6895, -0.6107]],

         [[-0.3987,  0.3658,  0.3610,  ...,  0.7149,  0.0373, -0.0760],
          [-0.9265,  1.6044, -1.0795,  ...,  1.7424,  0.7516,  0.4403],
          [-0.9426,  0.5687,  0.2647,  ...,  1.5837, -0.3063, -0.2318],
          [-0.1338, -1.0417, -1.2202,  ...,  1.9820, -0.5278, -0.7768],
          [-1.4452,  1.1149, -1.0638,  ...,  0.3779, -0.0381, -1.3205]],

         [[-0.7267, -0.0075,  0.4409,  ..., -0.0924,  0.0184, -0.0608],
          [ 0.6522, -0.2756,  0.4570,  ...,  0.1603, -1.1757,  0.9609],
          [ 0.2388, -0.8517,  0.7437,  ..., -0.2162, -0.6686,  1.3204],
          [ 0.1424,  0.7177,  1.3056,  ..., -0.3387, -0.9134,  1.7476],
          [ 0.2386, -0.3478,  2.2592,  ..., -0.1777, -0.0067,  1.8317]]]],
       grad_fn=<PermuteBackward0>), tensor([[[[ 7.4877e-02, -1.4604e-01, -1.8867e-01,  ..., -2.6349e-01,
            2.5296e-01, -1.3368e-01],
          [ 2.5747e+00,  5.2521e-01,  9.5710e-01,  ...,  3.1614e+00,
           -1.6083e+00,  4.1732e-01],
          [ 4.0702e-01, -1.8400e-01, -1.7714e-01,  ...,  3.5964e+00,
           -1.8526e+00,  1.6354e+00],
          [-1.4235e-01, -6.8709e-01, -8.0819e-01,  ...,  1.5049e+00,
           -2.2445e-01,  5.3625e-01],
          [-1.0486e+00,  2.9034e-01,  1.0123e+00,  ...,  2.0990e+00,
           -2.2622e+00,  1.7894e+00]],

         [[ 9.3830e-02, -2.8901e-02,  2.9833e-02,  ..., -1.8844e-02,
           -9.3671e-02,  1.6409e-01],
          [-4.0586e-01, -1.9469e-01, -6.7519e-01,  ...,  3.2622e-01,
            3.3009e-01, -6.6897e-01],
          [ 3.4676e-01, -8.4524e-02, -4.8182e-01,  ...,  5.3283e-01,
            1.7924e-01,  9.3449e-02],
          [ 4.3425e-01, -6.1975e-01, -3.5513e-01,  ..., -7.1465e-01,
           -3.6901e-01, -8.2377e-01],
          [ 1.1410e+00, -4.6568e-01,  6.3196e-01,  ..., -1.8152e-01,
           -2.5368e-01,  4.0097e-03]],

         [[ 2.1803e-03,  4.3629e-02, -4.4761e-02,  ...,  1.4167e-02,
            7.3670e-03,  5.8662e-02],
          [ 1.0830e+00,  3.5085e-01,  3.9491e-01,  ...,  5.0487e-01,
            3.6739e-01,  3.8893e-01],
          [ 1.1724e+00, -5.9721e-01, -4.6287e-01,  ...,  3.6452e-01,
           -6.5784e-01, -1.3521e-01],
          [-6.7131e-01,  1.9080e-01,  2.6190e+00,  ..., -1.9276e-01,
           -1.5709e+00,  1.9397e+00],
          [ 1.1348e+00, -1.4854e-01, -1.3176e+00,  ..., -5.3696e-01,
            1.0707e+00,  3.2988e-01]],

         ...,

         [[-1.3589e-02, -1.8782e-02,  1.1461e-01,  ...,  6.2486e-02,
           -9.7607e-03,  4.7317e-02],
          [ 5.6144e-03,  1.6483e-01,  4.0975e-01,  ...,  2.8187e-01,
           -8.9043e-03, -1.4201e-01],
          [-1.1171e+00,  9.5644e-01,  3.4411e-01,  ..., -3.7521e-01,
            3.9078e-01,  2.6438e-01],
          [-8.4454e-01, -1.2700e-01, -7.1156e-01,  ...,  8.3479e-01,
           -2.7073e+00, -2.6276e-01],
          [ 8.7597e-01, -5.9728e-01, -5.3489e-01,  ...,  3.8692e-01,
           -1.2302e+00,  7.4909e-01]],

         [[-1.8580e-01, -6.3994e-02,  7.1606e-02,  ..., -5.0592e-02,
            6.4941e-02, -9.8380e-02],
          [-8.4946e-01, -2.6254e-01, -6.7642e-02,  ..., -6.5549e-01,
            1.0445e+00,  4.7416e-01],
          [-2.7015e-01,  3.6563e-02,  3.2916e-01,  ..., -3.9337e-01,
            9.7853e-01,  1.6242e-02],
          [-1.5241e+00,  5.5751e-01, -1.0727e+00,  ...,  1.8536e+00,
            1.4632e+00, -1.7604e+00],
          [-1.4992e+00, -3.5960e-02,  7.0848e-01,  ..., -1.2286e-01,
            1.0103e-01, -1.2447e+00]],

         [[ 1.0095e-01, -1.4884e-01,  1.6019e-01,  ..., -1.4727e-01,
           -1.6222e-02, -1.8559e-01],
          [-7.9634e-01,  7.6121e-02,  3.0783e-01,  ..., -6.4954e-01,
           -4.3440e-01,  4.6446e-01],
          [-3.5833e-01, -1.4580e-01,  1.7373e-02,  ..., -5.7108e-01,
           -1.0706e-01,  7.0915e-02],
          [ 5.8874e-02,  1.1310e+00, -2.1959e-02,  ..., -2.8819e+00,
            1.1455e+00, -7.3015e-01],
          [-5.7424e-01,  5.9699e-01,  1.2147e+00,  ..., -5.5474e-01,
            1.3090e-01,  8.3539e-01]]]], grad_fn=<PermuteBackward0>))), hidden_states=None, attentions=None, cross_attentions=None)