Datasets:

Zyphra
/

Zyda

@@ -134,9 +134,13 @@ StarCoder: https://huggingface.co/datasets/bigcode/starcoderdata
 <!-- This section describes the data collection and processing process such as data selection criteria, filtering and normalization methods, tools and libraries used, etc. -->
 #### Personal and Sensitive Information

 <!-- This section describes the data collection and processing process such as data selection criteria, filtering and normalization methods, tools and libraries used, etc. -->
+Zyda was created using a two stage post-processing pipeline consisting of *filtering* and *deduplication*.
+For the filtering stage, we utilized a set of hand-crafted and tuned filters derived from a number of sources such as C4, RedPajama, and Gopher, in addition to our own filters.
+For the deduplication stage, we used minhash approximate deduplication. We deduplicated on 13-grams and used a minhash signature size of 128 and filtered out documents above a Jaccard similarity of 0.4.
+For full details on our data processing see the technical report.
 #### Personal and Sensitive Information