Update README.md

README.md

@@ -52,14 +52,15 @@ The dataset is divided into training and test splits:
 ### Keys (items)
 
 - `session_id`: "session_ge1nse2c",
-- `words`: Transcription corresponding to t
+- `words`: Transcription corresponding to the time stamp (start, end).
 - `start_time`: Start time in second.
 - `end_time`: End time in second.
 - `speaker`: Speaker tagging in string "speaker\<N\>"
 
 ### Source Datasets
 
-The dataset combines entries from various sources:
+`err_source_text`: This is the erroneous ASR-Diarization results to be fixed. Has dev, eval folders
+`ref_annotated_text`: This is the human annotated ground-truth for evaluation. Only dev split is included. 
 
 - **Development Sources**:
   - `dev`: 142 sessions
@@ -67,13 +68,10 @@ The dataset combines entries from various sources:
 - **Evaluation Sources**:
   - `eval`: 104 Sessions
 
-
-
 ## Access
 
 The dataset can be accessed and downloaded through the HuggingFace Datasets library.
 
-
 ## Evaluation
 
 This dataset can be evaluated by [MeetEval Software](https://github.com/fgnt/meeteval)