bram-w commited on
Commit
36c3ba1
1 Parent(s): 0eaa580

editing text

Browse files
Files changed (1) hide show
  1. app.py +6 -4
app.py CHANGED
@@ -42,7 +42,8 @@ def edict(x, source_text, edit_text,
42
 
43
  url = endpoint.url
44
  url = url + "/api/edit"
45
- headers = {
 
46
  "User-Agent": "EDICT HuggingFace Space",
47
  "Auth-Token": get_token(),
48
  }
@@ -93,15 +94,15 @@ for dog_i in [1, 2]:
93
  examples.append([f'square_ims/imagenet_dog_{dog_i}.jpg', 'A dog', f'A {breed}', 0.8, 3])
94
 
95
 
96
- description = 'A gradio demo for [EDICT](https://arxiv.org/abs/2211.12446, CVPR23)'
97
  # description = gr.Markdown(description)
98
 
99
  article = """
100
 
101
  ### Prompting Style
102
 
103
- As with many text-to-image methods, the prompting style of EDICT can make a big difference. When in doubt, experiment! Some guidance that we've found:
104
- * Parallel the Original and Edit description construction as much as possible. Inserting/editing single words often is enough to affect a change while maintaining a lot of the original structure
105
  * Words that will affect the entire setting (e.g. "A photo of " vs. "A painting of") can make a big difference. Playing around with them can help a lot
106
 
107
  ### Parameters
@@ -111,6 +112,7 @@ Both `edit_strength` and `guidance_scale` have similar properties qualitatively:
111
 
112
  Usually we find changing `edit_strength` to be enough, but feel free to play around (and report any interesting results)!
113
 
 
114
 
115
  Having difficulty coming up with a caption? Try [BLIP](https://huggingface.co/spaces/Salesforce/BLIP2) to automatically generate one!
116
 
 
42
 
43
  url = endpoint.url
44
  url = url + "/api/edit"
45
+ headers = {### Misc.
46
+
47
  "User-Agent": "EDICT HuggingFace Space",
48
  "Auth-Token": get_token(),
49
  }
 
94
  examples.append([f'square_ims/imagenet_dog_{dog_i}.jpg', 'A dog', f'A {breed}', 0.8, 3])
95
 
96
 
97
+ description = 'A gradio demo for [EDICT](https://arxiv.org/abs/2211.12446) (CVPR23)'
98
  # description = gr.Markdown(description)
99
 
100
  article = """
101
 
102
  ### Prompting Style
103
 
104
+ As with many text-to-image methods, the prompting style of EDICT can make a big difference. When in doubt, experiment! Some guidance:
105
+ * Parallel *Original Description* and *Edit Description* construction as much as possible. Inserting/editing single words often is enough to affect a change while maintaining a lot of the original structure
106
  * Words that will affect the entire setting (e.g. "A photo of " vs. "A painting of") can make a big difference. Playing around with them can help a lot
107
 
108
  ### Parameters
 
112
 
113
  Usually we find changing `edit_strength` to be enough, but feel free to play around (and report any interesting results)!
114
 
115
+ ### Misc.
116
 
117
  Having difficulty coming up with a caption? Try [BLIP](https://huggingface.co/spaces/Salesforce/BLIP2) to automatically generate one!
118