Update README.md

README.md

@@ -10,6 +10,10 @@ configs:
   data_files:
   - split: train
     path: minhash_deduped/**/*.parquet
+- config_name: quality_filtered
+  data_files:
+  - split: train
+    path: quality_filtered/**/*.parquet
 - config_name: matched
   data_files:
   - split: train
@@ -27,7 +31,8 @@ We train a 1.4B parameter language model through nanotron on 30 billion tokens t
 
 | Subset | Documents | Tokens | Description |
 |--------|-----------|--------|-------------|
-| `minhash_deduped` | 219.1M | 167.6B | Document-level MinHash deduplication only |
+| `quality_filtered` | 394.0M | 307.2B | Quality-filtered data before deduplication |
+| `minhash_deduped` | 219.1M | 167.6B | Document-level MinHash deduplication |
 | `matched` | 67.6M | 56.0B | Documents appearing in 2+ source datasets |
 
 The matched subset uses cross-dataset agreement as a signal for quality.
@@ -38,6 +43,7 @@ The matched subset uses cross-dataset agreement as a signal for quality.
 from datasets import load_dataset
 
 ds = load_dataset("AdaMLLab/TurMix", "minhash_deduped")
+ds = load_dataset("AdaMLLab/TurMix", "quality_filtered")
 ds = load_dataset("AdaMLLab/TurMix", "matched")
 ```