fffiloni lucaeyring commited on
Commit
d304d56
1 Parent(s): 925f03d

Fix example prompts (#3)

Browse files

- Fix example prompts (ca9b1c2dfe8f8cb65cda9846f8c4aaa50b14fadf)


Co-authored-by: Luca Eyring <lucaeyring@users.noreply.huggingface.co>

Files changed (1) hide show
  1. app.py +3 -3
app.py CHANGED
@@ -277,7 +277,7 @@ def combined_function(gallery_state, loaded_model_setup, prompt, chosen_model, s
277
  title="# ReNO: Enhancing One-step Text-to-Image Models through Reward-based Noise Optimization"
278
  description = "Enter a prompt to generate an image using ReNO. The method enhances text-to-image generation by optimizing \
279
  the initial noise using reward models as detailed in the paper. The demo uses a lower learning rate (2.5) compared to the paper's default (5.0) \
280
- for smoother trajectories - if you are looking for more dramatic changes, you can increase this value. You can also \
281
  adjust the reward weights to e.g. prioritize either prompt following (increase ImageReward) or aesthetic quality \
282
  (increase HPS/PickScore) based on your preferences.\n\nThe first time you load this demo, it will take a bit \
283
  to download and initialize the required model. Once loaded, each optimization run takes about 25-60 seconds."
@@ -344,10 +344,10 @@ with gr.Blocks(css=css, analytics_enabled=False) as demo:
344
  gr.Examples(
345
  examples = [
346
  "A red dog and a green cat",
 
347
  "A blue scooter is parked near a curb in front of a green vintage car",
348
  "A curious, orange fox and a fluffy, white rabbit, playing together in a lush, green meadow filled with yellow dandelions",
349
- "An orange chair to the right of a black airplane"
350
- "A toaster riding a bike",
351
  "A brain riding a rocketship towards the moon",
352
  ],
353
  inputs = [prompt]
 
277
  title="# ReNO: Enhancing One-step Text-to-Image Models through Reward-based Noise Optimization"
278
  description = "Enter a prompt to generate an image using ReNO. The method enhances text-to-image generation by optimizing \
279
  the initial noise using reward models as detailed in the paper. The demo uses a lower learning rate (2.5) compared to the paper's default (5.0) \
280
+ for smoother trajectories - if you are looking for more drastic changes, you can increase this value. You can also \
281
  adjust the reward weights to e.g. prioritize either prompt following (increase ImageReward) or aesthetic quality \
282
  (increase HPS/PickScore) based on your preferences.\n\nThe first time you load this demo, it will take a bit \
283
  to download and initialize the required model. Once loaded, each optimization run takes about 25-60 seconds."
 
344
  gr.Examples(
345
  examples = [
346
  "A red dog and a green cat",
347
+ "A toaster riding a bike",
348
  "A blue scooter is parked near a curb in front of a green vintage car",
349
  "A curious, orange fox and a fluffy, white rabbit, playing together in a lush, green meadow filled with yellow dandelions",
350
+ "An orange chair to the right of a black airplane",
 
351
  "A brain riding a rocketship towards the moon",
352
  ],
353
  inputs = [prompt]