Spaces:
Runtime error
Runtime error
Songwei Ge
commited on
Commit
•
ab7db7f
1
Parent(s):
b41079f
demo
Browse files- app.py +4 -22
- utils/attention_utils.py +11 -22
app.py
CHANGED
@@ -23,25 +23,6 @@ Instructions placeholder.
|
|
23 |
"""
|
24 |
|
25 |
|
26 |
-
example_instructions = [
|
27 |
-
"Make it a picasso painting",
|
28 |
-
"as if it were by modigliani",
|
29 |
-
"convert to a bronze statue",
|
30 |
-
"Turn it into an anime.",
|
31 |
-
"have it look like a graphic novel",
|
32 |
-
"make him gain weight",
|
33 |
-
"what would he look like bald?",
|
34 |
-
"Have him smile",
|
35 |
-
"Put him in a cocktail party.",
|
36 |
-
"move him at the beach.",
|
37 |
-
"add dramatic lighting",
|
38 |
-
"Convert to black and white",
|
39 |
-
"What if it were snowing?",
|
40 |
-
"Give him a leather jacket",
|
41 |
-
"Turn him into a cyborg!",
|
42 |
-
"make him wear a beanie",
|
43 |
-
]
|
44 |
-
|
45 |
def main():
|
46 |
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
47 |
model = RegionDiffusion(device)
|
@@ -90,9 +71,9 @@ def main():
|
|
90 |
height=height, width=width, num_inference_steps=steps,
|
91 |
guidance_scale=guidance_weight)
|
92 |
print('time lapses to get attention maps: %.4f' % (time.time()-begin_time))
|
93 |
-
color_obj_masks = get_token_maps(
|
94 |
model.attention_maps, run_dir, width//8, height//8, color_target_token_ids, seed)
|
95 |
-
model.masks = get_token_maps(
|
96 |
model.attention_maps, run_dir, width//8, height//8, region_target_token_ids, seed, base_tokens)
|
97 |
color_obj_masks = [transforms.functional.resize(color_obj_mask, (height, width),
|
98 |
interpolation=transforms.InterpolationMode.BICUBIC,
|
@@ -110,7 +91,8 @@ def main():
|
|
110 |
text_format_dict=text_format_dict)
|
111 |
print('time lapses to generate image from rich text: %.4f' %
|
112 |
(time.time()-begin_time))
|
113 |
-
|
|
|
114 |
|
115 |
with gr.Blocks() as demo:
|
116 |
gr.HTML("""<h1 style="font-weight: 900; margin-bottom: 7px;">Expressive Text-to-Image Generation with Rich Text</h1>
|
|
|
23 |
"""
|
24 |
|
25 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
26 |
def main():
|
27 |
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
28 |
model = RegionDiffusion(device)
|
|
|
71 |
height=height, width=width, num_inference_steps=steps,
|
72 |
guidance_scale=guidance_weight)
|
73 |
print('time lapses to get attention maps: %.4f' % (time.time()-begin_time))
|
74 |
+
color_obj_masks, _ = get_token_maps(
|
75 |
model.attention_maps, run_dir, width//8, height//8, color_target_token_ids, seed)
|
76 |
+
model.masks, token_maps = get_token_maps(
|
77 |
model.attention_maps, run_dir, width//8, height//8, region_target_token_ids, seed, base_tokens)
|
78 |
color_obj_masks = [transforms.functional.resize(color_obj_mask, (height, width),
|
79 |
interpolation=transforms.InterpolationMode.BICUBIC,
|
|
|
91 |
text_format_dict=text_format_dict)
|
92 |
print('time lapses to generate image from rich text: %.4f' %
|
93 |
(time.time()-begin_time))
|
94 |
+
cat_img = np.concatenate([plain_img[0], rich_img[0]], 1)
|
95 |
+
return [cat_img, token_maps]
|
96 |
|
97 |
with gr.Blocks() as demo:
|
98 |
gr.HTML("""<h1 style="font-weight: 900; margin-bottom: 7px;">Expressive Text-to-Image Generation with Rich Text</h1>
|
utils/attention_utils.py
CHANGED
@@ -76,15 +76,19 @@ def plot_attention_maps(atten_map_list, obj_tokens, save_dir, seed, tokens_vis=N
|
|
76 |
norm = mpl.colors.Normalize(vmin=vmin, vmax=vmax)
|
77 |
sm = plt.cm.ScalarMappable(cmap=cmap, norm=norm)
|
78 |
fig.colorbar(sm, cax=axs[-1])
|
|
|
|
|
|
|
|
|
79 |
|
80 |
fig.tight_layout()
|
81 |
plt.savefig(os.path.join(
|
82 |
save_dir, 'token_mapes_seed%d_%s.png' % (seed, atten_names[i])), dpi=100)
|
83 |
plt.close('all')
|
|
|
84 |
|
85 |
|
86 |
-
def get_token_maps(attention_maps, save_dir, width, height, obj_tokens, seed=0, tokens_vis=None
|
87 |
-
preprocess=False):
|
88 |
r"""Function to visualize attention maps.
|
89 |
Args:
|
90 |
save_dir (str): Path to save attention maps
|
@@ -177,23 +181,8 @@ def get_token_maps(attention_maps, save_dir, width, height, obj_tokens, seed=0,
|
|
177 |
attention_maps_averaged_normalized = [
|
178 |
attention_maps_averaged_normalized[i:i+1] for i in range(attention_maps_averaged_normalized.shape[0])]
|
179 |
|
180 |
-
|
181 |
-
|
182 |
-
|
183 |
-
|
184 |
-
|
185 |
-
attention_maps_averaged_eroded = [(torch.from_numpy(map).unsqueeze(
|
186 |
-
0)/255. > 0.8).float() for map in attention_maps_averaged_eroded]
|
187 |
-
attention_maps_averaged_eroded.append(
|
188 |
-
1 - torch.cat(attention_maps_averaged_eroded).sum(0, True))
|
189 |
-
plot_attention_maps([attention_maps_averaged, attention_maps_averaged_normalized,
|
190 |
-
attention_maps_averaged_eroded], obj_tokens, save_dir, seed, tokens_vis)
|
191 |
-
attention_maps_averaged_eroded = [attn_mask.unsqueeze(1).repeat(
|
192 |
-
[1, 4, 1, 1]).cuda() for attn_mask in attention_maps_averaged_eroded]
|
193 |
-
return attention_maps_averaged_eroded
|
194 |
-
else:
|
195 |
-
plot_attention_maps([attention_maps_averaged, attention_maps_averaged_normalized],
|
196 |
-
obj_tokens, save_dir, seed, tokens_vis)
|
197 |
-
attention_maps_averaged_normalized = [attn_mask.unsqueeze(1).repeat(
|
198 |
-
[1, 4, 1, 1]).cuda() for attn_mask in attention_maps_averaged_normalized]
|
199 |
-
return attention_maps_averaged_normalized
|
|
|
76 |
norm = mpl.colors.Normalize(vmin=vmin, vmax=vmax)
|
77 |
sm = plt.cm.ScalarMappable(cmap=cmap, norm=norm)
|
78 |
fig.colorbar(sm, cax=axs[-1])
|
79 |
+
canvas = fig.canvas
|
80 |
+
canvas.draw()
|
81 |
+
width, height = canvas.get_width_height()
|
82 |
+
img = np.frombuffer(canvas.tostring_rgb(), dtype='uint8').reshape((height, width, 3))
|
83 |
|
84 |
fig.tight_layout()
|
85 |
plt.savefig(os.path.join(
|
86 |
save_dir, 'token_mapes_seed%d_%s.png' % (seed, atten_names[i])), dpi=100)
|
87 |
plt.close('all')
|
88 |
+
return img
|
89 |
|
90 |
|
91 |
+
def get_token_maps(attention_maps, save_dir, width, height, obj_tokens, seed=0, tokens_vis=None):
|
|
|
92 |
r"""Function to visualize attention maps.
|
93 |
Args:
|
94 |
save_dir (str): Path to save attention maps
|
|
|
181 |
attention_maps_averaged_normalized = [
|
182 |
attention_maps_averaged_normalized[i:i+1] for i in range(attention_maps_averaged_normalized.shape[0])]
|
183 |
|
184 |
+
token_maps_vis = plot_attention_maps([attention_maps_averaged, attention_maps_averaged_normalized],
|
185 |
+
obj_tokens, save_dir, seed, tokens_vis)
|
186 |
+
attention_maps_averaged_normalized = [attn_mask.unsqueeze(1).repeat(
|
187 |
+
[1, 4, 1, 1]).cuda() for attn_mask in attention_maps_averaged_normalized]
|
188 |
+
return attention_maps_averaged_normalized, token_maps_vis
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|