Issue: Mismatch in Column Names for image_path and question
Description
There is a mismatch between the column names used in the code and the column names in the XLSX files under the Benchmark directory. The code currently refers to 'image_path' and 'question', but the XLSX files contain the column names 'Image' and 'Question'. This causes a KeyError when the script tries to access the columns.
Affected Code
The issue occurs between lines 107 and 109 in the llava_oeq_version.py script:
for index, row in df.iterrows():
image_file = os.path.join(image_base_path, row['image_path']) # Line 107-109
question = row['question']
Proposed Fix
To align with the actual column names in the XLSX files, the code should be updated as follows:
for index, row in df.iterrows():
image_file = os.path.join(image_base_path, row['Image']) # Corrected to 'Image'
question = row['Question'] # Corrected to 'Question'
Issue: Mismatch in Column Names for
image_path
andquestion
Description
There is a mismatch between the column names used in the code and the column names in the XLSX files under the
Benchmark
directory. The code currently refers to'image_path'
and'question'
, but the XLSX files contain the column names'Image'
and'Question'
. This causes aKeyError
when the script tries to access the columns.Affected Code
The issue occurs between lines 107 and 109 in the
llava_oeq_version.py
script:Proposed Fix
To align with the actual column names in the XLSX files, the code should be updated as follows: