fffiloni commited on
Commit
8ec125b
1 Parent(s): e42dad1

changes sentence extraction from kosmos2

Browse files
Files changed (1) hide show
  1. app.py +5 -20
app.py CHANGED
@@ -77,25 +77,10 @@ def get_caption(image_in):
77
  )
78
  print(f"KOSMOS2 RETURNS: {kosmos2_result}")
79
 
80
- with open(kosmos2_result[1], 'r') as f:
81
- data = json.load(f)
82
-
83
- reconstructed_sentence = []
84
- for sublist in data:
85
- reconstructed_sentence.append(sublist[0])
86
-
87
- full_sentence = ' '.join(reconstructed_sentence)
88
- #print(full_sentence)
89
-
90
- # Find the pattern matching the expected format ("Describe this image in detail:" followed by optional space and then the rest)...
91
- pattern = r'^Describe this image in detail:\s*(.*)$'
92
- # Apply the regex pattern to extract the description text.
93
- match = re.search(pattern, full_sentence)
94
- if match:
95
- description = match.group(1)
96
- print(description)
97
- else:
98
- print("Unable to locate valid description.")
99
 
100
  # Find the last occurrence of "."
101
  #last_period_index = full_sentence.rfind('.')
@@ -106,7 +91,7 @@ def get_caption(image_in):
106
  # print(truncated_caption)
107
  #print(f"\n—\nIMAGE CAPTION: {truncated_caption}")
108
 
109
- return description
110
 
111
  def get_caption_from_MD(image_in):
112
  client = Client("https://vikhyatk-moondream1.hf.space/")
 
77
  )
78
  print(f"KOSMOS2 RETURNS: {kosmos2_result}")
79
 
80
+ data = kosmos2_result[1]
81
+
82
+ # Extract and combine tokens starting from the second element
83
+ sentence = ''.join(item['token'] for item in data[1:])
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
84
 
85
  # Find the last occurrence of "."
86
  #last_period_index = full_sentence.rfind('.')
 
91
  # print(truncated_caption)
92
  #print(f"\n—\nIMAGE CAPTION: {truncated_caption}")
93
 
94
+ return sentence
95
 
96
  def get_caption_from_MD(image_in):
97
  client = Client("https://vikhyatk-moondream1.hf.space/")