ReleaseNVIDIANVIDIApublished Oct 20, 2025seen Jun 26

NVIDIA/recsys-examples v25.09

NVIDIA/recsys-examples

Open original ↗

Captured source

source ↗
GH
GitHub/github.com/NVIDIA/recsys-examples
NVIDIA/recsys-examples v25.09
Source ↗
published Oct 20, 2025seen Jun 26captured Aug 31http 200method plain

v25.09

Repository: NVIDIA/recsys-examples

Tag: v25.09

Published: 2025-10-20T10:00:01Z

Prerelease: no

Release notes:

What's Changed

Features & Enhancements

  • Dynamicemb prefetch integration by @JacoCheung in https://github.com/NVIDIA/recsys-examples/pull/181
  • Support distributed embedding dumping for dynamicemb by @z52527 @shijieliu in https://github.com/NVIDIA/recsys-examples/pull/120 https://github.com/NVIDIA/recsys-examples/pull/185
  • Add kernel fusion in HSTU block for inference, with KVCache fixes by @geoffreyQiu in https://github.com/NVIDIA/recsys-examples/pull/184
  • export hstu fp8 quant by @shijieliu in https://github.com/NVIDIA/recsys-examples/pull/168
  • Replace BatchedDynamicEmbeddingTables with BatchedDynamicEmbeddingTablesV2 by @jiashuy in https://github.com/NVIDIA/recsys-examples/pull/155

Bug Fixs

  • fix DynamicEmbDump - handle long strings in broadcast_string by @fshhr46 in https://github.com/NVIDIA/recsys-examples/pull/164
  • fix: consider mask when calc hstu attn flops by @shijieliu in https://github.com/NVIDIA/recsys-examples/pull/177
  • export fix hstu ima when num_candidates = seqlen by @shijieliu in https://github.com/NVIDIA/recsys-examples/pull/183

Misc

  • Make local hbm budget grow when num_embeddings grows. by @jiashuy in https://github.com/NVIDIA/recsys-examples/pull/156
  • Fix several errors for inference. by @geoffreyQiu in https://github.com/NVIDIA/recsys-examples/pull/167
  • Fix setup.py by @yiwenchen2025 in https://github.com/NVIDIA/recsys-examples/pull/169
  • Suppress mcore deps install by @JacoCheung in https://github.com/NVIDIA/recsys-examples/pull/170
  • dynamicemb clean BatchedDynamicEmbeddingTables by @jiashuy in https://github.com/NVIDIA/recsys-examples/pull/179
  • Update hstu layer benchmark doc by @JacoCheung in https://github.com/NVIDIA/recsys-examples/pull/171
  • Update dynamicemb's benchmark and example with README.md by @jiashuy in https://github.com/NVIDIA/recsys-examples/pull/188

New Contributors

  • @fshhr46 made their first contribution in https://github.com/NVIDIA/recsys-examples/pull/164

Full Changelog: https://github.com/NVIDIA/recsys-examples/compare/v25.08...v25.09