Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -290,6 +290,11 @@ with gr.Blocks(css=css) as demo:
|
|
| 290 |
CycleDiffusion (<a href="https://arxiv.org/abs/2210.05559">π Paper link</a> | <a href="https://huggingface.co/docs/diffusers/main/en/api/pipelines/cycle_diffusion">𧨠Pipeline doc</a>) is an image-to-image translation method that supports stochastic samplers for diffusion models. <br>
|
| 291 |
We also support the combination of CycleDiffusion and Cross Attention Control (CAC | <a href="https://arxiv.org/abs/2208.01626">π Paper link</a>). CAC is a technique to transfer the attention map from the source prompt to the target prompt. <br>
|
| 292 |
</p>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 293 |
<p>
|
| 294 |
<b>How to use:</b> <br>
|
| 295 |
1. Upload an image. <br>
|
|
@@ -304,7 +309,7 @@ with gr.Blocks(css=css) as demo:
|
|
| 304 |
<b>Notes:</b> <br>
|
| 305 |
1. CycleDiffusion is likely to fail when drastic changes are intended (e.g., changing a large black car to red). <br>
|
| 306 |
2. The value of strength can be set larger when CAC is used. <br>
|
| 307 |
-
3. If CAC type is "Replace", the source and target prompts should differ in only one token; otherwise, an error will be raised.
|
| 308 |
4. If CAC type is "Refine", the source prompt be a subsequence of the target prompt; otherwise, an error will be raised. <br>
|
| 309 |
</p>
|
| 310 |
<p>You can skip the queue using Colab: <a href="https://colab.research.google.com/gist/ChenWu98/0aa4fe7be80f6b45d3d055df9f14353a/copy-of-fine-tuned-diffusion-gradio.ipynb"><img data-canonical-src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab" src="https://colab.research.google.com/assets/colab-badge.svg"></a></p>
|
|
|
|
| 290 |
CycleDiffusion (<a href="https://arxiv.org/abs/2210.05559">π Paper link</a> | <a href="https://huggingface.co/docs/diffusers/main/en/api/pipelines/cycle_diffusion">𧨠Pipeline doc</a>) is an image-to-image translation method that supports stochastic samplers for diffusion models. <br>
|
| 291 |
We also support the combination of CycleDiffusion and Cross Attention Control (CAC | <a href="https://arxiv.org/abs/2208.01626">π Paper link</a>). CAC is a technique to transfer the attention map from the source prompt to the target prompt. <br>
|
| 292 |
</p>
|
| 293 |
+
<p>
|
| 294 |
+
<b>Quick start</b>: <br>
|
| 295 |
+
1. Click one row of Examples at the end of this page. It will fill all inputs needed. <br>
|
| 296 |
+
2. Click the "Edit" button. <br>
|
| 297 |
+
</p>
|
| 298 |
<p>
|
| 299 |
<b>How to use:</b> <br>
|
| 300 |
1. Upload an image. <br>
|
|
|
|
| 309 |
<b>Notes:</b> <br>
|
| 310 |
1. CycleDiffusion is likely to fail when drastic changes are intended (e.g., changing a large black car to red). <br>
|
| 311 |
2. The value of strength can be set larger when CAC is used. <br>
|
| 312 |
+
3. If CAC type is "Replace", the source and target prompts should differ in only one token; otherwise, an error will be raised. This is why we deliberately make some grammar mistakes in Examples.<br>
|
| 313 |
4. If CAC type is "Refine", the source prompt be a subsequence of the target prompt; otherwise, an error will be raised. <br>
|
| 314 |
</p>
|
| 315 |
<p>You can skip the queue using Colab: <a href="https://colab.research.google.com/gist/ChenWu98/0aa4fe7be80f6b45d3d055df9f14353a/copy-of-fine-tuned-diffusion-gradio.ipynb"><img data-canonical-src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab" src="https://colab.research.google.com/assets/colab-badge.svg"></a></p>
|