master

ascend-robot[Feature] Support full attention via Ring Attention under TND layout.

d76dbddd创建于 2025年11月25日历史提交

文件	最后提交记录	最后更新时间
__init__.py	!2246 refactor: context parallel refactor Merge pull request !2246 from YE ZHENYUAN/master	1 年前
context_parallel_kv_cache.py	!2796 mla supports ring-attention and tp Merge pull request !2796 from 吴炜强/mla-cp	8 个月前
ring_context_parallel.py	[Feature] Support full attention via Ring Attention under TND layout. Co-authored-by: zs-derrick1<1434012475@qq.com> # message auto-generated for no-merge-commit merge: !3021 merge master into master [Feature] Support full attention via Ring Attention under TND layout. Created-by: zs-derrick1 Commit-by: zs-derrick1 Merged-by: ascend-robot Description: [Feature] Support full attention via Ring Attention under TND layout. 添加新的AttentionWithCpTNDGeneral和TNDGeneralAttentionStrategy以支持TND layout在General 场景 下的Ring attention 验证结果： https://wiki.huawei.com/domains/149956/wiki/300791/WIKI202511048896138?title=_c3cd4b74 See merge request: Ascend/MindSpeed!3021	6 个月前