'pred_theta_mats' struct of rotate matrix data is row by row or column by column?

[weiweigt]

this data from output

'pred_theta_mats': tensor([[ 9.9878e-01, -4.8991e-02, 6.0638e-03, -4.9363e-02, -9.9217e-01, 1.1468e-01, 3.9805e-04, -1.1484e-01, -9.9338e-01, 9.9724e-01, 7.3746e-02, 8.9331e-03, -7.3243e-02, 9.9618e-01, -4.7464e-02, -1.2399e-02, 4.6679e-02, 9.9883e-01, 9.9725e-01, -7.0321e-02, -2.3572e-02, 6.7509e-02, 9.9227e-01, -1.0411e-01, 3.0711e-02, 1.0223e-01, 9.9429e-01, 9.9993e-01, -9.1336e-03, 7.8887e-03, 9.7961e-03, 9.9603e-01, -8.8491e-02, -7.0490e-03, 8.8562e-02, 9.9605e-01, 9.9667e-01, -1.9952e-02, -7.9075e-02, 3.4685e-04, 9.7064e-01, -2.4053e-01, 8.1553e-02, 2.3971e-01, 9.6741e-01, 9.9593e-01, 6.5499e-02, 6.1948e-02, -5.3998e-02, 9.8363e-01, -1.7191e-01, -7.2194e-02, 1.6787e-01, 9.8316e-01, 9.9995e-01, -8.8979e-04, 9.6254e-03, 8.8020e-04, 1.0000e+00, 9.0222e-04, -9.6243e-03, -8.9623e-04, 9.9995e-01, 9.7583e-01, 9.8794e-02, 1.9492e-01, -8.9844e-02, 9.9448e-01, -5.4253e-02, -1.9921e-01, 3.5429e-02, 9.7932e-01, 9.5518e-01, -1.3807e-01, -2.6186e-01, 1.1019e-01, 9.8683e-01, -1.1838e-01, 2.7476e-01, 8.4223e-02, 9.5782e-01, 9.9979e-01, 1.9870e-02, 4.8828e-03, -2.0016e-02, 9.9929e-01, 3.1854e-02, -4.2463e-03, -3.1945e-02, 9.9948e-01, 9.7423e-01, -1.2642e-01, 1.8680e-01, 1.2600e-01, 9.9193e-01, 1.4136e-02, -1.8708e-01, 9.7667e-03, 9.8229e-01, 9.4135e-01, 1.3354e-01, -3.0988e-01, -1.0419e-01, 9.8851e-01, 1.0949e-01, 3.2095e-01, -7.0778e-02, 9.4445e-01, 9.9051e-01, -1.2628e-01, -5.4184e-02, 1.2063e-01, 9.8793e-01, -9.7188e-02, 6.5803e-02, 8.9729e-02, 9.9379e-01, 9.9610e-01, -8.8154e-02, -2.8687e-03, 8.7510e-02, 9.8372e-01, 1.5698e-01, -1.1017e-02, -1.5662e-01, 9.8760e-01, 9.0254e-01, 3.7716e-02, 4.2895e-01, -5.4845e-02, 9.9811e-01, 2.7639e-02, -4.2710e-01, -4.8471e-02, 9.0290e-01, 9.9765e-01, 6.8491e-02, -2.9528e-03, -6.7717e-02, 9.9126e-01, 1.1322e-01, 1.0682e-02, -1.1276e-01, 9.9356e-01, 9.8918e-01, -8.4697e-02, 1.1979e-01, 3.5503e-02, 9.3045e-01, 3.6469e-01, -1.4235e-01, -3.5650e-01, 9.2339e-01, 7.7049e-01, -1.5181e-01, 6.1911e-01, -1.6263e-01, 8.9227e-01, 4.2119e-01, -6.1636e-01, -4.2520e-01, 6.6280e-01, 9.6830e-01, 2.4978e-01, 1.7518e-03, -2.3474e-01, 9.0755e-01, 3.4820e-01, 8.5384e-02, -3.3758e-01, 9.3742e-01, 9.9385e-01, 6.9942e-02, 8.5795e-02, -9.9011e-02, 9.0827e-01, 4.0651e-01, -4.9493e-02, -4.1250e-01, 9.0961e-01, 9.6430e-01, -1.2274e-01, -2.3466e-01, 1.1575e-01, 9.9233e-01, -4.3380e-02, 2.3818e-01, 1.4668e-02, 9.7111e-01, 9.5383e-01, 2.6260e-01, 1.4577e-01, -2.6036e-01, 9.6489e-01, -3.4538e-02, -1.4972e-01, -5.0103e-03, 9.8872e-01, 9.5834e-01, 2.7659e-01, 7.1255e-02, -2.8321e-01, 9.5255e-01, 1.1155e-01, -3.7022e-02, -1.2708e-01, 9.9120e-01, 9.9797e-01, 4.8428e-02, -4.1322e-02, -4.4938e-02, 9.9565e-01, 8.1576e-02, 4.5093e-02, -7.9552e-02, 9.9581e-01]], device='cuda:0', grad_fn=),