[API Compatibility] nn.Module.to_empty/linalg.det/nn.functional.gumbel_softmax/vstack/nn.functional.batch_norm/set_default_device/autograd.grad_mode.set_grad_enabled/Tensor.new_tensor/nn.modules.loss._Loss/nn.modules.utils._pair/hstack/Tensor.true_divide_/linalg.eigh/Tensor.H/nn.ELU/clamp_max/qr/logdet/linalg.cholesky/nn.functional.elu/nn.functional.prelu/expand_copy/linalg.qr/clamp_ Edit By AI Agent (#898)

* [API Compatibility] nn.Module.to_empty/linalg.det/nn.functional.gumbel_softmax/vstack/Tensor.gt_/Tensor.addcdiv/nn.functional.batch_norm/set_default_device/autograd.grad_mode.set_grad_enabled/Tensor.new_tensor/cuda.amp.GradScaler/nn.modules.loss._Loss/nn.modules.utils._pair Edit By AI Agent

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* [API Compatibility] skip test_linalg_qr mode=r cases due to Paddle single Tensor return

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* [API Compatibility] add 0-D test cases for Tensor.H and Tensor.mH

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix CI

---------

Co-authored-by: zhouwei25 <zhouwei25@baidu.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

Author: 3 people

paconvert/api_alias_mapping.json

@@ -123,7 +123,6 @@
   "torch.nn.modules.rnn.RNNCellBase": "torch.nn.RNNCellBase",
   "torch.nn.modules.sparse.Embedding": "torch.nn.Embedding",
   "torch.nn.parallel.DataParallel": "torch.nn.DataParallel",
-  "torch.nn.parallel.data_parallel.DataParallel": "torch.nn.DataParallel",
   "torch.nn.parallel.distributed.DistributedDataParallel": "torch.nn.parallel.DistributedDataParallel",
   "torch.nn.utils.clip_grad_norm": "torch.nn.utils.clip_grad_norm_",
   "torch.orgqr": "torch.linalg.householder_product",