[tests] add `torch.use_deterministic_algorithms` for XPU (#30774)

* add xpu check

* add marker

* add documentation

* update doc

* fix ci

* remove from global init

* fix

src/transformers/testing_utils.py

@@ -116,6 +116,7 @@ from .utils import (
     is_torch_bf16_available_on_device,
     is_torch_bf16_cpu_available,
     is_torch_bf16_gpu_available,
+    is_torch_deterministic,
     is_torch_fp16_available_on_device,
     is_torch_neuroncore_available,
     is_torch_npu_available,
@@ -943,6 +944,15 @@ def require_torch_bf16_cpu(test_case):
     )(test_case)
 
 
+def require_deterministic_for_xpu(test_case):
+    if is_torch_xpu_available():
+        return unittest.skipUnless(is_torch_deterministic(), "test requires torch to use deterministic algorithms")(
+            test_case
+        )
+    else:
+        return test_case
+
+
 def require_torch_tf32(test_case):
     """Decorator marking a test that requires Ampere or a newer GPU arch, cuda>=11 and torch>=1.7."""
     return unittest.skipUnless(

src/transformers/utils/__init__.py

@@ -188,6 +188,7 @@ from .import_utils import (
     is_torch_bf16_gpu_available,
     is_torch_compile_available,
     is_torch_cuda_available,
+    is_torch_deterministic,
     is_torch_fp16_available_on_device,
     is_torch_fx_available,
     is_torch_fx_proxy,

src/transformers/utils/import_utils.py

@@ -296,6 +296,18 @@ def is_torch_available():
     return _torch_available
 
 
+def is_torch_deterministic():
+    """
+    Check whether pytorch uses deterministic algorithms by looking if torch.set_deterministic_debug_mode() is set to 1 or 2"
+    """
+    import torch
+
+    if torch.get_deterministic_debug_mode() == 0:
+        return False
+    else:
+        return True
+
+
 def is_hqq_available():
     return _hqq_available
 

tests/models/encoder_decoder/test_modeling_encoder_decoder.py

@@ -18,7 +18,7 @@ import tempfile
 import unittest
 
 from transformers import is_torch_available, logging
-from transformers.testing_utils import CaptureLogger, require_torch, slow, torch_device
+from transformers.testing_utils import CaptureLogger, require_deterministic_for_xpu, require_torch, slow, torch_device
 
 from ...test_modeling_common import ids_tensor
 from ..bart.test_modeling_bart import BartStandaloneDecoderModelTester
@@ -639,6 +639,7 @@ class EncoderDecoderMixin:
         loss.backward()
 
     @slow
+    @require_deterministic_for_xpu
     def test_real_model_save_load_from_pretrained(self):
         model_2 = self.get_pretrained_model()
         model_2.to(torch_device)

tests/models/speech_encoder_decoder/test_modeling_speech_encoder_decoder.py

@@ -18,7 +18,7 @@ import tempfile
 import unittest
 
 from transformers import is_torch_available
-from transformers.testing_utils import require_torch, slow, torch_device
+from transformers.testing_utils import require_deterministic_for_xpu, require_torch, slow, torch_device
 
 from ...test_modeling_common import floats_tensor, ids_tensor, random_attention_mask
 from ..bert.test_modeling_bert import BertModelTester
@@ -422,6 +422,7 @@ class EncoderDecoderMixin:
         loss.backward()
 
     @slow
+    @require_deterministic_for_xpu
     def test_real_model_save_load_from_pretrained(self):
         model_2, inputs = self.get_pretrained_model_and_inputs()
         model_2.to(torch_device)
@@ -578,6 +579,7 @@ class Speech2TextBertModelTest(EncoderDecoderMixin, unittest.TestCase):
     def test_save_and_load_from_pretrained(self):
         pass
 
+    @require_deterministic_for_xpu
     # all published pretrained models are Speech2TextModel != Speech2TextEncoder
     def test_real_model_save_load_from_pretrained(self):
         pass

tests/models/speecht5/test_modeling_speecht5.py

@@ -22,6 +22,7 @@ import unittest
 from transformers import SpeechT5Config, SpeechT5HifiGanConfig
 from transformers.testing_utils import (
     is_torch_available,
+    require_deterministic_for_xpu,
     require_sentencepiece,
     require_tokenizers,
     require_torch,
@@ -1071,6 +1072,7 @@ class SpeechT5ForTextToSpeechIntegrationTests(unittest.TestCase):
             "Shape mismatch between generate_speech and generate methods.",
         )
 
+    @require_deterministic_for_xpu
     def test_one_to_many_generation(self):
         model = self.default_model
         processor = self.default_processor