Update README.md

README.md

@@ -114,16 +114,6 @@ After the shuffle completes, automated checks confirm:
 
 **HuggingFace Data Studio**: The parquet files were written without a page index, which prevents the HuggingFace Data Studio from serving random row previews without loading entire row groups. This does not affect programmatic consumption (PyArrow, pandas, DuckDB, etc.) — only the web-based Data Studio preview. A future re-serialization with `write_page_index=True` and smaller row-group sizes would resolve this.
 
-## Usage
-
-The output parquet files can be consumed by any parquet-compatible tool. For the Daisy training framework, feed them to the shard builder:
-
-```bash
-python data/build_smollm_fineweb_edu_dedup_shards.py \
-    --build fineweb-edu-dedup \
-    --local-parquet-dir data/fineweb-shuffled \
-    --fineweb-num-shards 2200
-```
 
 ## License