Hugging Face's logo

Spaces:
umichVision
/
virtex-redcaps
Runtime error

App Files Community
zamborg commited on
Commit
67da685
•
1 Parent(s): 7b528c6

i dummy pt 2 -> beam size != nucleus size

Browse files
Files changed (2) hide show
  1. app.py +1 -2
  2. config.yaml +2 -2
app.py CHANGED
@@ -59,8 +59,7 @@ else:
59
  with st.spinner("Generating Caption"):
60
  subreddit, caption = virtexModel.predict(image_dict)
61
  st.header("Predicted Caption:\n\n")
62
- st.subheader(f"Subreddit: {subreddit}\n")
63
- st.subheader(f"Caption: {caption}\n")
64
 
65
  image.close()
66
 
 
59
  with st.spinner("Generating Caption"):
60
  subreddit, caption = virtexModel.predict(image_dict)
61
  st.header("Predicted Caption:\n\n")
62
+ st.subheader(f"r/{subreddit}:\t{caption}\n")
 
63
 
64
  image.close()
65
 
config.yaml CHANGED
@@ -28,10 +28,10 @@ DATA:
28
  VOCAB_SIZE: 30000
29
  MODEL:
30
  DECODER:
31
- BEAM_SIZE: 0.5
32
  MAX_DECODING_STEPS: 30
33
  NAME: nucleus_sampling
34
- NUCLEUS_SIZE: 0.9
35
  LABEL_SMOOTHING: 0.1
36
  NAME: virtex_web
37
  TEXTUAL:
 
28
  VOCAB_SIZE: 30000
29
  MODEL:
30
  DECODER:
31
+ BEAM_SIZE: 5
32
  MAX_DECODING_STEPS: 30
33
  NAME: nucleus_sampling
34
+ NUCLEUS_SIZE: 0.5
35
  LABEL_SMOOTHING: 0.1
36
  NAME: virtex_web
37
  TEXTUAL: