Update app.py
Browse files
app.py
CHANGED
@@ -211,9 +211,6 @@ def generate_image(
|
|
211 |
image_1, caption_1,
|
212 |
image_2 = None, caption_2 = None,
|
213 |
image_3 = None, caption_3 = None,
|
214 |
-
use_id_1 = True,
|
215 |
-
use_id_2 = True,
|
216 |
-
use_id_3 = True,
|
217 |
num_inference_steps = 8,
|
218 |
cond_size = 256,
|
219 |
target_height = 768,
|
@@ -237,7 +234,7 @@ def generate_image(
|
|
237 |
|
238 |
images = [image_1, image_2, image_3]
|
239 |
captions = [caption_1, caption_2, caption_3]
|
240 |
-
idips_checkboxes = [
|
241 |
|
242 |
# βββ Fallback to VLM caption if any caption is empty βββ
|
243 |
for idx, (img, cap) in enumerate(zip(images, captions)):
|
@@ -643,24 +640,28 @@ if __name__ == "__main__":
|
|
643 |
"sample/woman2.jpg", "a woman",
|
644 |
"sample/dress.jpg", "a dress",
|
645 |
None, None,
|
|
|
646 |
],
|
647 |
[
|
648 |
"ENT1 wearing a tiny hat",
|
649 |
"sample/hamster.jpg", "a hamster",
|
650 |
None, None,
|
651 |
None, None,
|
|
|
652 |
],
|
653 |
[
|
654 |
"a drawing of ENT1 and ENT2 that the ENT1 is running alongside of a giant ENT2, in style of a comic book",
|
655 |
"sample/woman.jpg", "a woman",
|
656 |
"sample/hamster.jpg", "a hamster",
|
657 |
None, None,
|
|
|
658 |
],
|
659 |
[
|
660 |
"ENT1 with ENT2 holding ENT3",
|
661 |
"sample/sam.jpg", "a man",
|
662 |
"sample/hair.jpg", "curly hair",
|
663 |
"sample/can.jpg", "a can",
|
|
|
664 |
],
|
665 |
],
|
666 |
inputs=[
|
@@ -668,6 +669,7 @@ if __name__ == "__main__":
|
|
668 |
images[0], captions[0],
|
669 |
images[1], captions[1],
|
670 |
images[2], captions[2],
|
|
|
671 |
],
|
672 |
outputs=[output, final_text],
|
673 |
fn=generate_image,
|
@@ -682,10 +684,7 @@ if __name__ == "__main__":
|
|
682 |
prompt,
|
683 |
images[0], captions[0],
|
684 |
images[1], captions[1],
|
685 |
-
images[2], captions[2],
|
686 |
-
idip_checkboxes[0],
|
687 |
-
idip_checkboxes[1],
|
688 |
-
idip_checkboxes[2],
|
689 |
steps_slider,
|
690 |
cond_size,
|
691 |
target_height,
|
|
|
211 |
image_1, caption_1,
|
212 |
image_2 = None, caption_2 = None,
|
213 |
image_3 = None, caption_3 = None,
|
|
|
|
|
|
|
214 |
num_inference_steps = 8,
|
215 |
cond_size = 256,
|
216 |
target_height = 768,
|
|
|
234 |
|
235 |
images = [image_1, image_2, image_3]
|
236 |
captions = [caption_1, caption_2, caption_3]
|
237 |
+
idips_checkboxes = [True, True, True]
|
238 |
|
239 |
# βββ Fallback to VLM caption if any caption is empty βββ
|
240 |
for idx, (img, cap) in enumerate(zip(images, captions)):
|
|
|
640 |
"sample/woman2.jpg", "a woman",
|
641 |
"sample/dress.jpg", "a dress",
|
642 |
None, None,
|
643 |
+
8,
|
644 |
],
|
645 |
[
|
646 |
"ENT1 wearing a tiny hat",
|
647 |
"sample/hamster.jpg", "a hamster",
|
648 |
None, None,
|
649 |
None, None,
|
650 |
+
8,
|
651 |
],
|
652 |
[
|
653 |
"a drawing of ENT1 and ENT2 that the ENT1 is running alongside of a giant ENT2, in style of a comic book",
|
654 |
"sample/woman.jpg", "a woman",
|
655 |
"sample/hamster.jpg", "a hamster",
|
656 |
None, None,
|
657 |
+
8,
|
658 |
],
|
659 |
[
|
660 |
"ENT1 with ENT2 holding ENT3",
|
661 |
"sample/sam.jpg", "a man",
|
662 |
"sample/hair.jpg", "curly hair",
|
663 |
"sample/can.jpg", "a can",
|
664 |
+
24,
|
665 |
],
|
666 |
],
|
667 |
inputs=[
|
|
|
669 |
images[0], captions[0],
|
670 |
images[1], captions[1],
|
671 |
images[2], captions[2],
|
672 |
+
steps_slider,
|
673 |
],
|
674 |
outputs=[output, final_text],
|
675 |
fn=generate_image,
|
|
|
684 |
prompt,
|
685 |
images[0], captions[0],
|
686 |
images[1], captions[1],
|
687 |
+
images[2], captions[2],
|
|
|
|
|
|
|
688 |
steps_slider,
|
689 |
cond_size,
|
690 |
target_height,
|