Update diffusiondb_prompts_and_images_dataset.py

diffusiondb_prompts_and_images_dataset.py

@@ -52,27 +52,27 @@ Demo dataset for testing or showing image-text capabilities.
  ]
 
  def _generate_examples(self, images, parquet_file):
- df = pd.read_parquet(parquet_file)
- idx_counter = 0
- 
- # Create a dictionary from the dataframe for faster lookup
- df_dict = df.set_index('path').to_dict(orient='index')
- 
- for filepath, image in images:
- image_name = filepath.split('/')[-1]
- image_key = f"DiffusionDBImages/{image_name}"
+ df = pd.read_parquet(parquet_file)
+ idx_counter = 0
  
- # Check if the image_key exists in the dictionary
- if image_key in df_dict:
- row = df_dict[image_key]
- yield idx_counter, {
- "prompt": row['prompt'],
- "seed": row['seed'],
- "cfg": row['cfg'],
- "step": row['step'],
- "image": {"path": filepath, "bytes": image.read()},
- }
- idx_counter += 1
+ # Create a dictionary from the dataframe for faster lookup
+ df_dict = df.set_index('path').to_dict(orient='index')
+ 
+ for filepath, image in images:
+ image_name = filepath.split('/')[-1]
+ image_key = f"DiffusionDBImages/{image_name}"
+ 
+ # Check if the image_key exists in the dictionary
+ if image_key in df_dict:
+ row = df_dict[image_key]
+ yield idx_counter, {
+ "prompt": row['prompt'],
+ "seed": row['seed'],
+ "cfg": row['cfg'],
+ "step": row['step'],
+ "image": {"path": filepath, "bytes": image.read()},
+ }
+ idx_counter += 1
  
  # def _generate_examples(self, images, parquet_file):
  # df = pd.read_parquet(parquet_file)