Update README.md

README.md

@@ -4,16 +4,14 @@ license: gpl-3.0
 
 # SemanticGAN Dataset
 
-This repository contains the processed data for the SemanticGAN project. It is used as a data source for the training pipeline.
+This repository contains processed data used in the SemanticGAN experiments.
 
 ## Contents
+- kg_triples_ids.txt: approximately one million DBLP knowledge graph triples encoded as integer identifiers
+- kg_mappings.json: mapping between numeric identifiers and entity and relation labels
 
-- The file `kg_triples_ids.txt` consists of approximately one million DBLP knowledge graph triples encoded as integer identifiers.
-- The file `kg_mappings.json` stores the mapping between numeric identifiers and the corresponding entity and relation labels.
+## Notes
+The dataset is provided in processed form and stored separately from the model checkpoints.
 
-## Purpose
-
-The data is separated from the model repository to keep the model weights and training data organized. The GitHub Actions pipeline in the main repository pulls these files automatically during training.
-
-The model checkpoints and project description can be found at: 
-https://huggingface.co/erdemonal/SemanticGAN
+## Related resources
+- Model checkpoints: https://huggingface.co/erdemonal/SemanticGAN