Add support for Ascend NPU

Transformers is already supported on Ascend NPU. This PR allows users to easily use ChatGLM2 on NPU :-)

modeling_chatglm.py

@@ -29,7 +29,7 @@ from .configuration_chatglm import ChatGLMConfig
 
 # flags required to enable jit fusion kernels
 
-if sys.platform != 'darwin':
+if sys.platform != 'darwin' and torch.cuda.is_available():
     torch._C._jit_set_profiling_mode(False)
     torch._C._jit_set_profiling_executor(False)
     torch._C._jit_override_can_fuse_on_cpu(True)