[Tests] Fix mps+generator fast tests (#1230)

* [Tests] Fix mps+generator fast tests

* mps for Euler

* retry

* warmup issue again?

* fix reproducible initial noise

* Revert "fix reproducible initial noise"

This reverts commit f300d05cb9782ed320064a0c58577a32d4139e6d.

* fix reproducible initial noise

* fix device

.github/workflows/pr_tests.yml

@@ -136,7 +136,7 @@ jobs:
     - name: Run fast PyTorch tests on M1 (MPS)
       shell: arch -arch arm64 bash {0}
       run: |
-        ${CONDA_RUN} python -m pytest -n 1 -s -v --make-reports=tests_torch_mps tests/
+        ${CONDA_RUN} python -m pytest -n 0 -s -v --make-reports=tests_torch_mps tests/
 
     - name: Failure short reports
       if: ${{ failure() }}

src/diffusers/pipelines/ddim/pipeline_ddim.py

@@ -78,7 +78,7 @@ class DDIMPipeline(DiffusionPipeline):
         if generator is not None and generator.device.type != self.device.type and self.device.type != "mps":
             message = (
                 f"The `generator` device is `{generator.device}` and does not match the pipeline "
-                f"device `{self.device}`, so the `generator` will be set to `None`. "
+                f"device `{self.device}`, so the `generator` will be ignored. "
                 f'Please use `generator=torch.Generator(device="{self.device}")` instead.'
             )
             deprecate(
@@ -89,11 +89,13 @@ class DDIMPipeline(DiffusionPipeline):
             generator = None
 
         # Sample gaussian noise to begin loop
-        image = torch.randn(
-            (batch_size, self.unet.in_channels, self.unet.sample_size, self.unet.sample_size),
-            generator=generator,
-            device=self.device,
-        )
+        image_shape = (batch_size, self.unet.in_channels, self.unet.sample_size, self.unet.sample_size)
+        if self.device.type == "mps":
+            # randn does not work reproducibly on mps
+            image = torch.randn(image_shape, generator=generator)
+            image = image.to(self.device)
+        else:
+            image = torch.randn(image_shape, generator=generator, device=self.device)
 
         # set step values
         self.scheduler.set_timesteps(num_inference_steps)

src/diffusers/pipelines/ddpm/pipeline_ddpm.py

@@ -83,7 +83,7 @@ class DDPMPipeline(DiffusionPipeline):
         if generator is not None and generator.device.type != self.device.type and self.device.type != "mps":
             message = (
                 f"The `generator` device is `{generator.device}` and does not match the pipeline "
-                f"device `{self.device}`, so the `generator` will be set to `None`. "
+                f"device `{self.device}`, so the `generator` will be ignored. "
                 f'Please use `torch.Generator(device="{self.device}")` instead.'
             )
             deprecate(
@@ -94,11 +94,13 @@ class DDPMPipeline(DiffusionPipeline):
             generator = None
 
         # Sample gaussian noise to begin loop
-        image = torch.randn(
-            (batch_size, self.unet.in_channels, self.unet.sample_size, self.unet.sample_size),
-            generator=generator,
-            device=self.device,
-        )
+        image_shape = (batch_size, self.unet.in_channels, self.unet.sample_size, self.unet.sample_size)
+        if self.device.type == "mps":
+            # randn does not work reproducibly on mps
+            image = torch.randn(image_shape, generator=generator)
+            image = image.to(self.device)
+        else:
+            image = torch.randn(image_shape, generator=generator, device=self.device)
 
         # set step values
         self.scheduler.set_timesteps(num_inference_steps)

tests/pipelines/ddpm/test_ddpm.py

@@ -81,10 +81,14 @@ class DDPMPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         if torch_device == "mps":
             _ = ddpm(num_inference_steps=1)
 
+        if torch_device == "mps":
+            # device type MPS is not supported for torch.Generator() api.
+            generator = torch.manual_seed(0)
+        else:
             generator = torch.Generator(device=torch_device).manual_seed(0)
         image = ddpm(generator=generator, num_inference_steps=2, output_type="numpy").images
 
-        generator = torch.Generator(device=torch_device).manual_seed(0)
+        generator = generator.manual_seed(0)
         image_eps = ddpm(generator=generator, num_inference_steps=2, output_type="numpy", predict_epsilon=False)[0]
 
         image_slice = image[0, -3:, -3:, -1]

tests/test_scheduler.py

@@ -1281,7 +1281,11 @@ class EulerDiscreteSchedulerTest(SchedulerCommonTest):
 
         scheduler.set_timesteps(self.num_inference_steps)
 
-        generator = torch.Generator(torch_device).manual_seed(0)
+        if torch_device == "mps":
+            # device type MPS is not supported for torch.Generator() api.
+            generator = torch.manual_seed(0)
+        else:
+            generator = torch.Generator(device=torch_device).manual_seed(0)
 
         model = self.dummy_model()
         sample = self.dummy_sample_deter * scheduler.init_noise_sigma
@@ -1308,7 +1312,11 @@ class EulerDiscreteSchedulerTest(SchedulerCommonTest):
 
         scheduler.set_timesteps(self.num_inference_steps, device=torch_device)
 
-        generator = torch.Generator(torch_device).manual_seed(0)
+        if torch_device == "mps":
+            # device type MPS is not supported for torch.Generator() api.
+            generator = torch.manual_seed(0)
+        else:
+            generator = torch.Generator(device=torch_device).manual_seed(0)
 
         model = self.dummy_model()
         sample = self.dummy_sample_deter * scheduler.init_noise_sigma
@@ -1364,6 +1372,10 @@ class EulerAncestralDiscreteSchedulerTest(SchedulerCommonTest):
 
         scheduler.set_timesteps(self.num_inference_steps)
 
+        if torch_device == "mps":
+            # device type MPS is not supported for torch.Generator() api.
+            generator = torch.manual_seed(0)
+        else:
             generator = torch.Generator(device=torch_device).manual_seed(0)
 
         model = self.dummy_model()
@@ -1381,7 +1393,7 @@ class EulerAncestralDiscreteSchedulerTest(SchedulerCommonTest):
         result_sum = torch.sum(torch.abs(sample))
         result_mean = torch.mean(torch.abs(sample))
 
-        if str(torch_device).startswith("cpu"):
+        if torch_device in ["cpu", "mps"]:
             assert abs(result_sum.item() - 152.3192) < 1e-2
             assert abs(result_mean.item() - 0.1983) < 1e-3
         else:
@@ -1396,6 +1408,10 @@ class EulerAncestralDiscreteSchedulerTest(SchedulerCommonTest):
 
         scheduler.set_timesteps(self.num_inference_steps, device=torch_device)
 
+        if torch_device == "mps":
+            # device type MPS is not supported for torch.Generator() api.
+            generator = torch.manual_seed(0)
+        else:
             generator = torch.Generator(device=torch_device).manual_seed(0)
 
         model = self.dummy_model()