chunked/content_aware_chunking/_glossary/chunk_137.txt

Shards gets 
processed separately and in parallel on different GPUs and the results are synced at the end of the processing step.