Fix quality

tests/models/whisper/test_modeling_flax_whisper.py

@@ -34,11 +34,11 @@ if is_datasets_available():
     from datasets import load_dataset
 
 if is_flax_available():
-    import numpy as np
-
     import jax
+    import numpy as np
     from flax.core.frozen_dict import unfreeze
     from flax.traverse_util import flatten_dict
+
     from transformers import (
         FLAX_MODEL_MAPPING,
         FlaxWhisperForConditionalGeneration,

tests/models/whisper/test_modeling_whisper.py

@@ -51,6 +51,7 @@ if is_torch_available():
 
 if is_flax_available():
     import jax.numpy as jnp
+
     from transformers.modeling_flax_pytorch_utils import (
         convert_pytorch_state_dict_to_flax,
         load_flax_weights_in_pytorch_model,