From Chunks to Blocks: Accelerating Uploads and Downloads on the Hub

In three linesHugging Face optimizes uploads and downloads on the Hub by replacing chunks with blocks. This architecture reduces latency and improves stability for large file transfers.

Summary generated by Claude — human-verified