diff --git a/.github/workflows/pr_tests.yml b/.github/workflows/pr_tests.yml
index f63d4ffda464..c25aa888f459 100644
--- a/.github/workflows/pr_tests.yml
+++ b/.github/workflows/pr_tests.yml
@@ -21,7 +21,7 @@ jobs:
     runs-on: [ self-hosted, docker-gpu ]
     container:
       image: python:3.7
-      options: --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+      options: --shm-size "16gb" --ipc host -v /mnt/hf_cache:/mnt/cache/
 
     steps:
     - name: Checkout diffusers
diff --git a/.github/workflows/push_tests.yml b/.github/workflows/push_tests.yml
index 3db6814e071d..3e4a81c91c01 100644
--- a/.github/workflows/push_tests.yml
+++ b/.github/workflows/push_tests.yml
@@ -15,14 +15,10 @@ env:
 jobs:
   run_tests_single_gpu:
     name: Diffusers tests
-    strategy:
-      fail-fast: false
-      matrix:
-        machine_type: [ single-gpu ]
-    runs-on: [ self-hosted, docker-gpu, '${{ matrix.machine_type }}' ]
+    runs-on: [ self-hosted, docker-gpu, single-gpu ]
     container:
       image: nvcr.io/nvidia/pytorch:22.07-py3
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/hf_cache:/mnt/cache
 
     steps:
     - name: Checkout diffusers
@@ -66,14 +62,10 @@ jobs:
 
   run_examples_single_gpu:
     name: Examples tests
-    strategy:
-      fail-fast: false
-      matrix:
-        machine_type: [ single-gpu ]
-    runs-on: [ self-hosted, docker-gpu, '${{ matrix.machine_type }}' ]
+    runs-on: [ self-hosted, docker-gpu, single-gpu ]
     container:
       image: nvcr.io/nvidia/pytorch:22.07-py3
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/hf_cache:/mnt/cache
 
     steps:
     - name: Checkout diffusers
diff --git a/setup.py b/setup.py
index ff5f14564487..2e87d75960ca 100644
--- a/setup.py
+++ b/setup.py
@@ -92,7 +92,7 @@
     "jaxlib>=0.1.65,<=0.3.6",
     "modelcards>=0.1.4",
     "numpy",
-    "onnxruntime-gpu",
+    "onnxruntime",
     "pytest",
     "pytest-timeout",
     "pytest-xdist",
@@ -178,7 +178,7 @@ def run(self):
 extras["training"] = deps_list("accelerate", "datasets", "tensorboard", "modelcards")
 extras["test"] = deps_list(
     "datasets",
-    "onnxruntime-gpu",
+    "onnxruntime",
     "pytest",
     "pytest-timeout",
     "pytest-xdist",
diff --git a/src/diffusers/dependency_versions_table.py b/src/diffusers/dependency_versions_table.py
index 82ca5dbb6f56..367b5c57a262 100644
--- a/src/diffusers/dependency_versions_table.py
+++ b/src/diffusers/dependency_versions_table.py
@@ -17,7 +17,7 @@
     "jaxlib": "jaxlib>=0.1.65,<=0.3.6",
     "modelcards": "modelcards>=0.1.4",
     "numpy": "numpy",
-    "onnxruntime-gpu": "onnxruntime-gpu",
+    "onnxruntime": "onnxruntime",
     "pytest": "pytest",
     "pytest-timeout": "pytest-timeout",
     "pytest-xdist": "pytest-xdist",
diff --git a/tests/test_pipelines.py b/tests/test_pipelines.py
index d0d78171378e..cde652f3b828 100644
--- a/tests/test_pipelines.py
+++ b/tests/test_pipelines.py
@@ -1422,18 +1422,18 @@ def test_stable_diffusion_inpaint_pipeline_k_lms(self):
     @slow
     def test_stable_diffusion_onnx(self):
         sd_pipe = StableDiffusionOnnxPipeline.from_pretrained(
-            "CompVis/stable-diffusion-v1-4", revision="onnx", provider="CUDAExecutionProvider", use_auth_token=True
+            "CompVis/stable-diffusion-v1-4", revision="onnx", provider="CPUExecutionProvider", use_auth_token=True
         )
 
         prompt = "A painting of a squirrel eating a burger"
         np.random.seed(0)
-        output = sd_pipe([prompt], guidance_scale=6.0, num_inference_steps=20, output_type="np")
+        output = sd_pipe([prompt], guidance_scale=6.0, num_inference_steps=5, output_type="np")
         image = output.images
 
         image_slice = image[0, -3:, -3:, -1]
 
         assert image.shape == (1, 512, 512, 3)
-        expected_slice = np.array([0.0385, 0.0252, 0.0234, 0.0287, 0.0358, 0.0287, 0.0276, 0.0235, 0.0010])
+        expected_slice = np.array([0.3602, 0.3688, 0.3652, 0.3895, 0.3782, 0.3747, 0.3927, 0.4241, 0.4327])
         assert np.abs(image_slice.flatten() - expected_slice).max() < 1e-3
 
     @slow
@@ -1592,7 +1592,7 @@ def test_callback_fn(step: int, timestep: int, latents: np.ndarray) -> None:
                 assert latents.shape == (1, 4, 64, 64)
                 latents_slice = latents[0, -3:, -3:, -1]
                 expected_slice = np.array(
-                    [-0.6254, -0.2742, -1.0710, 0.2296, -1.1683, 0.6913, -2.0605, -0.0682, 0.9700]
+                    [-0.5950, -0.3039, -1.1672, 0.1594, -1.1572, 0.6719, -1.9712, -0.0403, 0.9592]
                 )
                 assert np.abs(latents_slice.flatten() - expected_slice).max() < 1e-3
 
@@ -1606,6 +1606,6 @@ def test_callback_fn(step: int, timestep: int, latents: np.ndarray) -> None:
         prompt = "Andromeda galaxy in a bottle"
 
         np.random.seed(0)
-        pipe(prompt=prompt, num_inference_steps=50, guidance_scale=7.5, callback=test_callback_fn, callback_steps=1)
+        pipe(prompt=prompt, num_inference_steps=5, guidance_scale=7.5, callback=test_callback_fn, callback_steps=1)
         assert test_callback_fn.has_been_called
-        assert number_of_steps == 51
+        assert number_of_steps == 6