Spaces:
Runtime error
Runtime error
Update preprocessor.py
Browse files- preprocessor.py +5 -4
preprocessor.py
CHANGED
|
@@ -23,7 +23,7 @@ from transformers import pipeline
|
|
| 23 |
|
| 24 |
class DepthEstimator:
|
| 25 |
def __init__(self):
|
| 26 |
-
self.model = pipeline("
|
| 27 |
|
| 28 |
def __call__(self, image: np.ndarray, **kwargs) -> PIL.Image.Image:
|
| 29 |
detect_resolution = kwargs.pop("detect_resolution", 512)
|
|
@@ -55,7 +55,8 @@ def resize_image(input_image, resolution, interpolation=None):
|
|
| 55 |
|
| 56 |
|
| 57 |
class Preprocessor:
|
| 58 |
-
MODEL_ID = "condition/ckpts"
|
|
|
|
| 59 |
|
| 60 |
def __init__(self):
|
| 61 |
self.model = None
|
|
@@ -73,8 +74,8 @@ class Preprocessor:
|
|
| 73 |
elif name == "Canny":
|
| 74 |
self.model = CannyDetector()
|
| 75 |
elif name == "Depth":
|
| 76 |
-
|
| 77 |
-
self.model = MidasDetector.from_pretrained(self.MODEL_ID)
|
| 78 |
else:
|
| 79 |
raise ValueError
|
| 80 |
torch.cuda.empty_cache()
|
|
|
|
| 23 |
|
| 24 |
class DepthEstimator:
|
| 25 |
def __init__(self):
|
| 26 |
+
self.model = pipeline("depth-estimation")
|
| 27 |
|
| 28 |
def __call__(self, image: np.ndarray, **kwargs) -> PIL.Image.Image:
|
| 29 |
detect_resolution = kwargs.pop("detect_resolution", 512)
|
|
|
|
| 55 |
|
| 56 |
|
| 57 |
class Preprocessor:
|
| 58 |
+
# MODEL_ID = "condition/ckpts"
|
| 59 |
+
MODEL_ID = "lllyasviel/Annotators"
|
| 60 |
|
| 61 |
def __init__(self):
|
| 62 |
self.model = None
|
|
|
|
| 74 |
elif name == "Canny":
|
| 75 |
self.model = CannyDetector()
|
| 76 |
elif name == "Depth":
|
| 77 |
+
self.model = DepthEstimator()
|
| 78 |
+
# self.model = MidasDetector.from_pretrained(self.MODEL_ID)
|
| 79 |
else:
|
| 80 |
raise ValueError
|
| 81 |
torch.cuda.empty_cache()
|