NVIDIA/recsys-examples v26.01

v26.01

Repository: NVIDIA/recsys-examples

Tag: v26.01

Published: 2026-02-13T02:41:48Z

Prerelease: no

Release notes:

What's Changed

Features & Enhancements

• HSTU KV Cache Manager V2 by @geoffreyQiu in https://github.com/NVIDIA/recsys-examples/pull/251
• workload balancer and datasets folder refactor by @JacoCheung in https://github.com/NVIDIA/recsys-examples/pull/275
• Fea unify pooling to dynamic embedding table by @shijieliu in https://github.com/NVIDIA/recsys-examples/pull/301
• Optimize EmbeddingBagCollection preliminarily by @jiashuy in https://github.com/NVIDIA/recsys-examples/pull/268
• refactor unique to stateless op by @shijieliu in https://github.com/NVIDIA/recsys-examples/pull/290
• Optimize dedup indices and segmented unique by @shijieliu in https://github.com/NVIDIA/recsys-examples/pull/293
• optimize backward local_reduce, use fwd unique results by @shijieliu in https://github.com/NVIDIA/recsys-examples/pull/299

Bug Fixes

• Fix devel build failure by @JacoCheung in https://github.com/NVIDIA/recsys-examples/pull/288
• Fix wrong evicted values when insert failed/busy in insert_and_evict. by @jiashuy in https://github.com/NVIDIA/recsys-examples/pull/284
• Fix test_jagged_tensor import bug by @JacoCheung in https://github.com/NVIDIA/recsys-examples/pull/289
• Fix KVCounter initialization bug. by @z52527 in https://github.com/NVIDIA/recsys-examples/pull/282
• Fix mcore version in training readme by @JacoCheung in https://github.com/NVIDIA/recsys-examples/pull/286
• Fix issue related to empty batch by @jiashuy in https://github.com/NVIDIA/recsys-examples/pull/271
• Fix issue #272: dump/load score consistency in STEP mode by @ShaobinChen-AH in https://github.com/NVIDIA/recsys-examples/pull/298

Misc

• Pull new triton hstu kernel by @JacoCheung in https://github.com/NVIDIA/recsys-examples/pull/291
• Add README for embedding pooling. by @z52527 in https://github.com/NVIDIA/recsys-examples/pull/270
• rename wrong module name by @gameofdimension in https://github.com/NVIDIA/recsys-examples/pull/278
• [CI] split dynamicemb tests by @shijieliu in https://github.com/NVIDIA/recsys-examples/pull/273
• Update triton version by @JacoCheung in https://github.com/NVIDIA/recsys-examples/pull/287
• release v26.01 by @shijieliu in https://github.com/NVIDIA/recsys-examples/pull/309

New Contributors

• @gameofdimension made their first contribution in https://github.com/NVIDIA/recsys-examples/pull/278
• @ShaobinChen-AH made their first contribution in https://github.com/NVIDIA/recsys-examples/pull/298

Full Changelog: https://github.com/NVIDIA/recsys-examples/compare/v25.12...v26.01