[misc] refactor launch API and tensor constructor (#5666)

* [misc] remove config arg from initialize

* [misc] remove old tensor contrusctor

* [plugin] add npu support for ddp

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* [devops] fix doc test ci

* [test] fix test launch

* [doc] update launch doc

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

colossalai/booster/plugin/torch_ddp_plugin.py

@@ -10,6 +10,7 @@ from colossalai.checkpoint_io import CheckpointIO, GeneralCheckpointIO
 from colossalai.cluster import DistCoordinator
 from colossalai.interface import ModelWrapper, OptimizerWrapper
 from colossalai.quantization import BnbQuantizationConfig, quantize_model
+from colossalai.utils import get_current_device
 
 from .dp_plugin_base import DPPluginBase
 
@@ -203,7 +204,7 @@ class TorchDDPPlugin(DPPluginBase):
         return True
 
     def supported_devices(self) -> List[str]:
-        return ["cuda"]
+        return ["cuda", "npu"]
 
     def configure(
         self,
@@ -214,7 +215,7 @@ class TorchDDPPlugin(DPPluginBase):
         lr_scheduler: Optional[LRScheduler] = None,
     ) -> Tuple[nn.Module, OptimizerWrapper, Callable, DataLoader, LRScheduler]:
         # cast model to cuda
-        model = model.cuda()
+        model = model.to(get_current_device())
 
         # convert model to sync bn
         model = nn.SyncBatchNorm.convert_sync_batchnorm(model, None)