[pipeline] Fix str device issue (#24396)

* fix str device issue

* fixup

* adapt from suggestions

* forward contrib credits from suggestions

* better fix

* added backward compatibility for older PT versions

* final fixes

* oops

* Attempting something with less branching.

---------

Co-authored-by: amyeroberts <amyeroberts@users.noreply.github.com>
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>

src/transformers/pipelines/base.py

@@ -903,9 +903,10 @@ class Pipeline(_ScikitCompat):
                 yield
         else:
             if self.device.type == "cuda":
-                torch.cuda.set_device(self.device)
-
-            yield
+                with torch.cuda.device(self.device):
+                    yield
+            else:
+                yield
 
     def ensure_tensor_on_device(self, **inputs):
         """

tests/pipelines/test_pipelines_common.py

@@ -46,6 +46,7 @@ from transformers.testing_utils import (
     require_tensorflow_probability,
     require_tf,
     require_torch,
+    require_torch_gpu,
     require_torch_or_tf,
     slow,
 )
@@ -542,6 +543,20 @@ class PipelineUtilsTest(unittest.TestCase):
         gc.collect()
         torch.cuda.empty_cache()
 
+    @slow
+    @require_torch
+    @require_torch_gpu
+    def test_pipeline_cuda(self):
+        pipe = pipeline("text-generation", device="cuda")
+        _ = pipe("Hello")
+
+    @slow
+    @require_torch
+    @require_torch_gpu
+    def test_pipeline_cuda_indexed(self):
+        pipe = pipeline("text-generation", device="cuda:0")
+        _ = pipe("Hello")
+
     @slow
     @require_tf
     @require_tensorflow_probability