mohdelgaar commited on
Commit
fbe88e2
β€’
1 Parent(s): 54ba470
Files changed (1) hide show
  1. app.py +9 -10
app.py CHANGED
@@ -228,7 +228,7 @@ title = """
228
  The model can generate diverse paraphrases of a given sentence, each adjusted to maintain consistent meaning while varying
229
  in linguistic complexity according to the desired level.</p>
230
  <p style="font-size:1.2em;">It is important to note that not all index combinations are feasible (e.g., a sentence of "length" 5 with 10 "unique words").
231
- To ensure high-quality outputs, our approach interpolates the embeddings of linguistic indices to identify the closest,
232
  achievable set of indices for the given target.</p>
233
  """
234
 
@@ -247,19 +247,18 @@ paraphrase generation). Then, click "Generate."
247
  **Manual Linguistic Control**: Select this option to manually control the linguistic complexity of the
248
  generated text. We provided a set of tools for manual adjustments of the desired linguistic complexity of
249
  the target sentence. These tools enable the user to extract linguistic indices from a given sentence,
250
- generate a random (yet coherent) set of linguistic indices, and add or remove noise from the indices.
251
  These tools are designed for experimental use and require the user to possess linguistic expertise for
252
  effective input of linguistic indices. To use these tools, select "Tools to assist in setting linguistic
253
  indices." Once indices are entered, click "Generate."
254
 
255
 
256
- Second, you may select to use exact or approximate computation of linguistic indices (used in mode (2) and
257
- in quality control of the genration). Approximate computation is significantly faster.
258
 
259
- Third, you may view the intermediate sentences of the quality control process by selecting the checkbox.
260
 
261
- Fourth, you may try out some examples by clicking on "Examples...". Examples consist of a source sentences,
262
- the indices of the source sentences, and a sample set of target linguistic indices.
263
 
264
  Please make your choice below.
265
 
@@ -291,13 +290,13 @@ body {
291
 
292
  .top-separator {
293
  width: 100%;
294
- height: 3px; /* Adjust the height for boldness */
295
  background-color: #000; /* Adjust the color as needed */
296
  margin-top: 20px; /* Adjust the margin as needed */
297
  }
298
  .bottom-separator {
299
  width: 100%;
300
- height: 3px; /* Adjust the height for boldness */
301
  background-color: #000; /* Adjust the color as needed */
302
  margin-bottom: 20px; /* Adjust the margin as needed */
303
  }
@@ -319,7 +318,7 @@ with gr.Blocks(
319
  with gr.Group(elem_id='mode'):
320
  mode = gr.Radio(
321
  value='Randomized Paraphrase Generation',
322
- label='How would you like to use this system?',
323
  type="index",
324
  choices=['πŸ”„ Randomized Paraphrase Generation',
325
  'βš–οΈ Complexity-Matched Paraphrasing',
 
228
  The model can generate diverse paraphrases of a given sentence, each adjusted to maintain consistent meaning while varying
229
  in linguistic complexity according to the desired level.</p>
230
  <p style="font-size:1.2em;">It is important to note that not all index combinations are feasible (e.g., a sentence of "length" 5 with 10 "unique words").
231
+ To ensure high-quality outputs, our approach iteratively adjusts the generated text to match the closest, yet coherent
232
  achievable set of indices for the given target.</p>
233
  """
234
 
 
247
  **Manual Linguistic Control**: Select this option to manually control the linguistic complexity of the
248
  generated text. We provided a set of tools for manual adjustments of the desired linguistic complexity of
249
  the target sentence. These tools enable the user to extract linguistic indices from a given sentence,
250
+ generate a random (yet coherent) set of linguistic indices, and add or subtract to them.
251
  These tools are designed for experimental use and require the user to possess linguistic expertise for
252
  effective input of linguistic indices. To use these tools, select "Tools to assist in setting linguistic
253
  indices." Once indices are entered, click "Generate."
254
 
255
 
256
+ Second, you may select to use exact or approximate computation of linguistic indices. Approximate computation is significantly faster.
 
257
 
258
+ Third, you may view the intermediate sentences of the quality control process by selecting the checkbox under "Advanced Options".
259
 
260
+ Fourth, you may try out some examples by clicking on "Examples...". Examples consist of a source sentences
261
+ and a sample set of target linguistic indices.
262
 
263
  Please make your choice below.
264
 
 
290
 
291
  .top-separator {
292
  width: 100%;
293
+ height: 4px; /* Adjust the height for boldness */
294
  background-color: #000; /* Adjust the color as needed */
295
  margin-top: 20px; /* Adjust the margin as needed */
296
  }
297
  .bottom-separator {
298
  width: 100%;
299
+ height: 4px; /* Adjust the height for boldness */
300
  background-color: #000; /* Adjust the color as needed */
301
  margin-bottom: 20px; /* Adjust the margin as needed */
302
  }
 
318
  with gr.Group(elem_id='mode'):
319
  mode = gr.Radio(
320
  value='Randomized Paraphrase Generation',
321
+ label='Operation Modes',
322
  type="index",
323
  choices=['πŸ”„ Randomized Paraphrase Generation',
324
  'βš–οΈ Complexity-Matched Paraphrasing',