[LoRA] allow big CUDA tests to run properly for LoRA (and others)

tests/pipelines/mochi/test_mochi.py

@@ -17,13 +17,16 @@
 import unittest
 
 import numpy as np
+import pytest
 import torch
 from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLMochi, FlowMatchEulerDiscreteScheduler, MochiPipeline, MochiTransformer3DModel
 from diffusers.utils.testing_utils import (
     enable_full_determinism,
+    nightly,
     numpy_cosine_similarity_distance,
+    require_big_gpu_with_torch_cuda,
     require_torch_gpu,
     slow,
     torch_device,
@@ -261,7 +264,10 @@ def test_vae_tiling(self, expected_diff_max: float = 0.2):
 
 
 @slow
+@nightly
 @require_torch_gpu
+@require_big_gpu_with_torch_cuda
+@pytest.mark.big_gpu_with_torch_cuda
 class MochiPipelineIntegrationTests(unittest.TestCase):
     prompt = "A painting of a squirrel eating a burger."
 
@@ -275,7 +281,7 @@ def tearDown(self):
         gc.collect()
         torch.cuda.empty_cache()
 
-    def test_cogvideox(self):
+    def test_mochi(self):
         generator = torch.Generator("cpu").manual_seed(0)
 
         pipe = MochiPipeline.from_pretrained("genmo/mochi-1-preview", torch_dtype=torch.float16)
@@ -293,7 +299,7 @@ def test_cogvideox(self):
         ).frames
 
         video = videos[0]
-        expected_video = torch.randn(1, 16, 480, 848, 3).numpy()
+        expected_video = torch.randn(1, 19, 480, 848, 3).numpy()
 
         max_diff = numpy_cosine_similarity_distance(video, expected_video)
         assert max_diff < 1e-3, f"Max diff is too high. got {video}"