Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
@@ -43,8 +43,15 @@ pipe = StableDiffusionXLFillPipeline.from_pretrained(
|
|
43 |
|
44 |
pipe.scheduler = TCDScheduler.from_config(pipe.scheduler.config)
|
45 |
|
46 |
-
# 텍스트
|
47 |
pipe.text_encoder = pipe.text_encoder.to("cuda", dtype=torch.float16)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
48 |
|
49 |
def can_expand(source_width, source_height, target_width, target_height, alignment):
|
50 |
"""Checks if the image can be expanded based on the alignment."""
|
|
|
43 |
|
44 |
pipe.scheduler = TCDScheduler.from_config(pipe.scheduler.config)
|
45 |
|
46 |
+
# 텍스트 인코더를 float16으로 강제 변환
|
47 |
pipe.text_encoder = pipe.text_encoder.to("cuda", dtype=torch.float16)
|
48 |
+
# 추가: text_projection의 forward를 오버라이딩하여 입력이 float16이 아니면 half로 캐스팅
|
49 |
+
original_text_projection_forward = pipe.text_encoder.text_projection.forward
|
50 |
+
def fixed_text_projection_forward(x):
|
51 |
+
if x.dtype != torch.float16:
|
52 |
+
x = x.half()
|
53 |
+
return original_text_projection_forward(x)
|
54 |
+
pipe.text_encoder.text_projection.forward = fixed_text_projection_forward
|
55 |
|
56 |
def can_expand(source_width, source_height, target_width, target_height, alignment):
|
57 |
"""Checks if the image can be expanded based on the alignment."""
|