Tom Aarsen commited on
Commit
760f88c
·
1 Parent(s): d86454c

Turn info section into HTML for Spaces; worked with Markdown only locally

Browse files
Files changed (1) hide show
  1. app.py +7 -6
app.py CHANGED
@@ -170,7 +170,7 @@ def search(
170
 
171
 
172
  css = """
173
- #cards-container {
174
  --block-padding: 0px;
175
  }
176
  """
@@ -178,15 +178,16 @@ css = """
178
  with gr.Blocks(title="Quantized Retrieval") as demo:
179
  with gr.Row():
180
  with gr.Column(scale=3):
181
- gr.Markdown(
182
  """
183
  <div style='border: 1px solid var(--border-color-primary, #e0e0e0); border-radius: var(--block-radius); padding: 12px 14px; background-color: var(--block-background-fill, transparent);'>
184
 
185
  <h1 style='margin-top: 0;'>Quantized Retrieval - Binary Search with Scalar (int8) Rescoring</h1>
186
 
187
- This demo showcases retrieval using [quantized embeddings](https://huggingface.co/blog/embedding-quantization) on a CPU. The corpus consists of [41 million texts](https://huggingface.co/datasets/sentence-transformers/quantized-retrieval-data) from Wikipedia articles.
188
  </div>
189
- """
 
190
  )
191
  with gr.Accordion("Click to learn about the retrieval process", open=False):
192
  gr.Markdown(
@@ -212,7 +213,7 @@ Feel free to check out the [code for this demo](https://huggingface.co/spaces/se
212
  Notes:
213
  - The approximate search index (a binary Inverted File Index (IVF)) is in beta and has not been trained with a lot of data.
214
  """
215
- )
216
  query = gr.Textbox(
217
  label="Query for Wikipedia articles",
218
  placeholder="Enter a query to search for relevant texts from Wikipedia.",
@@ -243,7 +244,7 @@ Notes:
243
 
244
  with gr.Row():
245
  with gr.Column(scale=3):
246
- cards = gr.HTML(label="Results", elem_id="cards-container")
247
  with gr.Column(scale=1):
248
  summary = gr.Markdown(label="Search Summary")
249
 
 
170
 
171
 
172
  css = """
173
+ .no-pad-container {
174
  --block-padding: 0px;
175
  }
176
  """
 
178
  with gr.Blocks(title="Quantized Retrieval") as demo:
179
  with gr.Row():
180
  with gr.Column(scale=3):
181
+ gr.HTML(
182
  """
183
  <div style='border: 1px solid var(--border-color-primary, #e0e0e0); border-radius: var(--block-radius); padding: 12px 14px; background-color: var(--block-background-fill, transparent);'>
184
 
185
  <h1 style='margin-top: 0;'>Quantized Retrieval - Binary Search with Scalar (int8) Rescoring</h1>
186
 
187
+ This demo showcases retrieval using<a href="https://huggingface.co/blog/embedding-quantization" style="padding-left: 0.5ch; padding-right: 0.5ch;">quantized embeddings</a>on a CPU. The corpus consists of<a href="https://huggingface.co/datasets/sentence-transformers/quantized-retrieval-data" style="padding-left: 0.5ch; padding-right: 0.5ch;">41 million texts</a>from Wikipedia articles.
188
  </div>
189
+ """,
190
+ elem_classes="no-pad-container",
191
  )
192
  with gr.Accordion("Click to learn about the retrieval process", open=False):
193
  gr.Markdown(
 
213
  Notes:
214
  - The approximate search index (a binary Inverted File Index (IVF)) is in beta and has not been trained with a lot of data.
215
  """
216
+ )
217
  query = gr.Textbox(
218
  label="Query for Wikipedia articles",
219
  placeholder="Enter a query to search for relevant texts from Wikipedia.",
 
244
 
245
  with gr.Row():
246
  with gr.Column(scale=3):
247
+ cards = gr.HTML(label="Results", elem_classes="no-pad-container")
248
  with gr.Column(scale=1):
249
  summary = gr.Markdown(label="Search Summary")
250