Tools · Updated 6 Oct, 10:09 pm IST
Transformers v5.19 adds EmbeddingGemma2; Polars 2.0 ships SQL and out-of-core
Why it matters for readers: If you follow ML and data tooling, these updates add a new multimodal embedding and make Polars more capable for SQL and large-data tasks.
- Transformers v5.19.0 introduces EmbeddingGemma2, a multimodal embedding that encodes text, images, audio and video into a shared 768-dimensional vector space.2
- EmbeddingGemma2 supports Matryoshka Representation Learning allowing embeddings to be truncated to 512, 256, or 128 dimensions and offers configurable visual and video token budgets.2
- Polars 2.0 ships initial out-of-core (spill-to-disk) support and a set of core performance improvements.1
- Polars 2.0 treats SQL as a first-class citizen with expanded SQL coverage and engine/optimizer improvements, and adds a new Map dtype and stricter dtype semantics.1
- Polars 2.0 authors report benchmark comparisons on TPC-H and TPC-DS-derived data against DuckDB and DataFusion as part of demonstrating performance gains.1
Get a brief like this every morning
Uzha reads hundreds of sources and gives you the stories that matter for your work, with every source linked. Free.
Get startedHow this story developed
6 Oct, 05:29 pm · Hacker News (100+ points)
Release of Polars 2.0
6 Oct, 10:09 pm · Transformers releases
Release v5.19.0