采用地平线stereonetplus网络训练自制双目数据集发生问题

1.芯片： J5

2.OE包版本：v1.1.68

3.问题描述:(1)  运行时有警告 :

2025-04-22 16:18:30,121 WARNING [hash.py:218] Node[0] Don not found hash value in name of /open_explorer/configs/disparity_pred/officaldownloadWeight/float-checkpoint-best.pth.tar, will skip check hash... 2025-04-22 16:18:30,257 WARNING [checkpoint.py:67] Node[0] module. is not at the beginning of state dict

警告2：2025-04-22 16:18:34,691 WARNING: Force duplicate shared conv-bn is disabled by default as of version 1.9.0. If you still need this feature before version 1.11.0 to load old checkpoints or for other reasons, please set `horizon_plugin_pytorch.qat_mode.tricks.fx_force_duplicate_shared_convbn = True` to enable it. However, please note that this feature will be removed in version 1.11.0. `aidisdk` dependency is not available. WARNING:root:init `TorchModulePatch` failed, caused by 'Required dependencies is not available: ModuleNotFoundError: No module named 'hatbc'. ', will set `patcher=None` `aidisdk` dependency is not available.

警告3：

1......./usr/local/lib/python3.8/dist-packages/hat/data/collates/collates.py:67: UserWarning: An output with one or more elements was resized since it had shape [8], which does not match the required output shape [2, 4]. This behavior is deprecated, and in a future PyTorch release outputs will not be resized unless they have zero elements. You can explicitly reuse an out tensor t by resizing it, inplace, to zero elements with t.resize_(0). (Triggered internally at ../aten/src/ATen/native/Resize.cpp:17.) return torch.stack(batch, 0, out=out)

2........./usr/local/lib/python3.8/dist-packages/hat/data/collates/collates.py:67: UserWarning: An output with one or more elements was resized since it had shape [4177920], which does not match the required output shape [2, 1088, 1920]. This behavior is deprecated, and in a future PyTorch release outputs will not be resized unless they have zero elements. You can explicitly reuse an out tensor t by resizing it, inplace, to zero elements with t.resize_(0). (Triggered internally at ../aten/src/ATen/native/Resize.cpp:17.) return torch.stack(batch, 0, out=out)发现输出的预测图像为(1088,1920)跟label的尺寸（1080，1920）有所差别，麻烦大佬帮忙讲解应该在哪里修改把