master

ascend-robotdocs: bm

f1b3d84d创建于 1月23日历史提交

文件	最后提交记录	最后更新时间
megatron_moe	docs(fboverlap): add usage of nopp Co-authored-by: clc2025<chenlucong@huawei.com> # message auto-generated for no-merge-commit merge: !3084 merge master into master docs(fboverlap): add usage of nopp Created-by: clc2025 Commit-by: clc2025 Merged-by: ascend-robot Description: add usage of nopp and modify some problems of format See merge request: Ascend/MindSpeed!3084	8 个月前
Automatic_Parallelism.md	Modify non-standard descriptions in the document Co-authored-by: clc2025<chenlucong@huawei.com> # message auto-generated for no-merge-commit merge: !2933 merge master into master Modify non-standard descriptions in the document Created-by: clc2025 Commit-by: clc2025 Merged-by: ascend-robot Description: DTS2025101630022 修改文档中的不规范描述 See merge request: Ascend/MindSpeed!2933	10 个月前
LayerZeRO.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
activation-function-recompute.md	Modify non-standard descriptions in the document Co-authored-by: clc2025<chenlucong@huawei.com> # message auto-generated for no-merge-commit merge: !2933 merge master into master Modify non-standard descriptions in the document Created-by: clc2025 Commit-by: clc2025 Merged-by: ascend-robot Description: DTS2025101630022 修改文档中的不规范描述 See merge request: Ascend/MindSpeed!2933	10 个月前
aiqos.md	AI Qos Feature Co-authored-by: EX_mitsuX<yangjie409@h-partners.com> Co-authored-by: ascend-robot<zhongyuanke@huawei.com> Co-authored-by: Klayyy<wanglei886@h-partners.com> Co-authored-by: LinMingZhe<linmingzhe3@huawei.com> Co-authored-by: clc2025<chenlucong@huawei.com> Co-authored-by: JialiZheng1<jializheng@huawei.com> Co-authored-by: MissingPompeii<guohao120@huawei.com> Co-authored-by: wangjinyi6<wangjinyi6@huawei.com> Co-authored-by: Klayyy<Klayyy@noreply.gitcode.com> Co-authored-by: w30064656<wangzhuangzhuang8@h-partners.com> Co-authored-by: xubin787<mark19980312@126.com> Co-authored-by: libaokui<libaokui@huawei.com> Co-authored-by: LinShua<707894133@qq.com> Co-authored-by: guihaowen666<guihaowen@huawei.com> Co-authored-by: ChenDonYY<caichendong2@huawei.com> Co-authored-by: wuweiqiang24<wuweiqiang11@huawei.com> Co-authored-by: Muuyo<koimuu@163.com> Co-authored-by: ybwang19<1605891897@qq.com> Co-authored-by: lmztju<limingzhao3@h-partners.com> Co-authored-by: xuyujun<xuyujun5@hisilicon.com> Co-authored-by: Jia_Austin<dengjia6@huawei.com> # message auto-generated for no-merge-commit merge: !3161 merge qos into master feat:AI QoS Feature Created-by: Klayyy Commit-by: Klayyy;Jia_Austin;libaokui;EX_mitsuX;wuweiqiang24;Muuyo;xuyujun;lmztju;wangjinyi6;ascend-robot;ybwang19;xubin787;ChenDonYY;guihaowen666;clc2025;LinShua;w30064656;MissingPompeii;JialiZheng1;LinMingZhe Merged-by: ascend-robot Description: AI QoS Feature See merge request: Ascend/MindSpeed!3161	7 个月前
alibi.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
async-ddp-param-gather.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
async-ddp.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
async-log-allreduce.md	Modify non-standard descriptions in the document Co-authored-by: clc2025<chenlucong@huawei.com> # message auto-generated for no-merge-commit merge: !2933 merge master into master Modify non-standard descriptions in the document Created-by: clc2025 Commit-by: clc2025 Merged-by: ascend-robot Description: DTS2025101630022 修改文档中的不规范描述 See merge request: Ascend/MindSpeed!2933	10 个月前
auto_settings.md	fixbug for auto_settings Co-authored-by: wuweiqiang24<wuweiqiang11@huawei.com> # message auto-generated for no-merge-commit merge: !3003 merge auto_settings into master fixbug for auto_settings Created-by: wuweiqiang24 Commit-by: wuweiqiang24 Merged-by: ascend-robot Description: 1. fixbug for auto_settings 2. add readme See merge request: Ascend/MindSpeed!3003	8 个月前
automated-pipeline.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
balanced_moe.md	docs: bm Co-authored-by: Jia_Austin<dengjia6@huawei.com> # message auto-generated for no-merge-commit merge: !3226 merge bm into master docs: bm Created-by: Jia_Austin Commit-by: Jia_Austin Merged-by: ascend-robot Description: docs: bm See merge request: Ascend/MindSpeed!3226	7 个月前
communication-over-computation.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
compress-tensor.md	feat: 内存压缩特性迭代升级 Co-authored-by: NingGuangyou<ningguangyou@h-partners.com> # message auto-generated for no-merge-commit merge: !3173 merge master into master feat: 内存压缩特性迭代升级 Created-by: NingGuangyou Commit-by: NingGuangyou Merged-by: ascend-robot Description: 本次PR将原首节点MLP模块激活值压缩特性迭代为各节点按transformer layer激活值压缩及AdamW一二阶动量压缩，是保留原特性及使用方法不变的前提下增加了新的功能。详情可参考readme。 See merge request: Ascend/MindSpeed!3173	7 个月前
context_parallelism_kv_cache.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
conv3d_sequence_paralle.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
custom_fsdp.md	update doc of custom-fsdp Co-authored-by: clc2025<chenlucong@huawei.com> # message auto-generated for no-merge-commit merge: !3098 merge master into master update doc of custom-fsdp Created-by: clc2025 Commit-by: clc2025 Merged-by: ascend-robot Description: update doc of custom-fsdp See merge request: Ascend/MindSpeed!3098	8 个月前
data-parallel.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
dist_ckpt.md	Modify non-standard descriptions in the document Co-authored-by: clc2025<chenlucong@huawei.com> # message auto-generated for no-merge-commit merge: !2933 merge master into master Modify non-standard descriptions in the document Created-by: clc2025 Commit-by: clc2025 Merged-by: ascend-robot Description: DTS2025101630022 修改文档中的不规范描述 See merge request: Ascend/MindSpeed!2933	10 个月前
distributed-optimizer.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
double-ring.md	add readme for double ring Co-authored-by: wuweiqiang24<wuweiqiang11@huawei.com> # message auto-generated for no-merge-commit merge: !3118 merge add_readme_double_ring into master add readme for double ring Created-by: wuweiqiang24 Commit-by: wuweiqiang24 Merged-by: ascend-robot Description: 增加约束描述 See merge request: Ascend/MindSpeed!3118	8 个月前
dualpipev.md	fix【问题】use_custom_fsdp特性和dualpipev不兼容 Co-authored-by: lizhi<lizhi166@huawei.com> # message auto-generated for no-merge-commit merge: !3080 merge bug-master into master fix【问题】use_custom_fsdp特性和dualpipev不兼容 Created-by: liz_hw Commit-by: lizhi Merged-by: ascend-robot Description: fix【问题】use_custom_fsdp特性和dualpipev不兼容 See merge request: Ascend/MindSpeed!3080	8 个月前
eod-reset.md	!2673 【feat.】支持动态padding功能 Merge pull request !2673 from yuqi/dynamic_padding	1 年前
flash-attention.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
fsdp2.md	[Bugfix]Init with Meta device when using fsdp2 Co-authored-by: zs-derrick1<1434012475@qq.com> # message auto-generated for no-merge-commit merge: merge master into master [Bugfix]Init with Meta device when using fsdp2 Created-by: zs-derrick1 Commit-by: zs-derrick1 Merged-by: ascend-robot Description: [Bugfix]Init with Meta device when using fsdp2 See merge request: Ascend/MindSpeed!2853	11 个月前
fused_ema_adamw_optimizer.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
fusion-attn-v2.md	!2376 【Readme update!!!】资料修正 Merge pull request !2376 from yangjie/master	1 年前
generate-mask.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
hccl-group-buffer-set.md	Modify non-standard descriptions in the document Co-authored-by: clc2025<chenlucong@huawei.com> # message auto-generated for no-merge-commit merge: !2933 merge master into master Modify non-standard descriptions in the document Created-by: clc2025 Commit-by: clc2025 Merged-by: ascend-robot Description: DTS2025101630022 修改文档中的不规范描述 See merge request: Ascend/MindSpeed!2933	10 个月前
hccl-replace-gloo.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
hybrid-context-parallel.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
img.png	!2821 add mla readme Merge pull request !2821 from 吴炜强/mla-readme	11 个月前
kvallgather-context-parallel.md	feat: kvallgather_cp_algo Co-authored-by: wuweiqiang24<wuweiqiang11@huawei.com> # message auto-generated for no-merge-commit merge: !3104 merge add_allgather_cp into master feat: kvallgather_cp_algo Created-by: wuweiqiang24 Commit-by: wuweiqiang24 Merged-by: ascend-robot Description: 1. 增加TE中CPStrategyFactory、BaseCPStrategy标准化接口 2. 增加KVAllGatherCPStrategy策略 3. 增加KVAllGatherCP Readme文档 4. 增加KVAllGatherCP UT测试用例 精度已对齐，并进行了性能测试： https://wiki.huawei.com/domains/137239/wiki/268925/WIKI202512179461897 See merge request: Ascend/MindSpeed!3104	8 个月前
low-precision-optimizer.md	低精度优化器 增加reademe Co-authored-by: w30064656<wangzhuangzhuang8@h-partners.com> # message auto-generated for no-merge-commit merge: !3067 merge master into master 低精度优化器 增加reademe Created-by: w30064656 Commit-by: w30064656 Merged-by: ascend-robot Description: 增加reademe 修复bug See merge request: Ascend/MindSpeed!3067	8 个月前
mc2.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
moe-token-permute-and-unpermute.md	!2778 fix: MoEAlltoAllSeqOverLapDispatcher and moe_permute_fusion Merge pull request !2778 from 邓佳/core_r0.12.1_fix_permute	1 年前
moe_expert_load_balance_placeemnt.md	增加MOE专家负载均衡功能 Co-authored-by: zhanggaolu2<252028123@qq.com> # message auto-generated for no-merge-commit merge: !2845 merge expert_loadbalance2master into master 增加MOE专家负载均衡功能 Created-by: zhanggaolu2 Commit-by: zhanggaolu2 Merged-by: ascend-robot Description: 增加MOE专家负载均衡功能 See merge request: Ascend/MindSpeed!2845	10 个月前
multi-head-latent-attention.md	!2821 add mla readme Merge pull request !2821 from 吴炜强/mla-readme	11 个月前
multi_parameter_pipeline.md	!2479 docs:modify docs for multiparameter Merge pull request !2479 from liurong1995/feature_docs	1 年前
multi_parameter_pipeline_and_variable_seq_lengths.md	!2479 docs:modify docs for multiparameter Merge pull request !2479 from liurong1995/feature_docs	1 年前
noop-layers.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
norm-recompute.md	Modify non-standard descriptions in the document Co-authored-by: clc2025<chenlucong@huawei.com> # message auto-generated for no-merge-commit merge: !2933 merge master into master Modify non-standard descriptions in the document Created-by: clc2025 Commit-by: clc2025 Merged-by: ascend-robot Description: DTS2025101630022 修改文档中的不规范描述 See merge request: Ascend/MindSpeed!2933	10 个月前
npu_datadump.md	feat: MindSpeed datadump!!! Co-authored-by: EX_mitsu<yangjie409@h-partners.com> # message auto-generated for no-merge-commit merge: !3166 merge master into master feat: MindSpeed datadump!!! Created-by: EX_mitsuX Commit-by: EX_mitsuX;EX_mitsu Merged-by: ascend-robot Description: 支持利用mstt dump整网精度数据。 See merge request: Ascend/MindSpeed!3166	7 个月前
npu_deterministic.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
npu_matmul_add.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
ops_flops_cal.md	revise readme of op-cal-tflops Co-authored-by: wuweiqiang24<wuweiqiang11@huawei.com> # message auto-generated for no-merge-commit merge: merge add_tflops_readme into master revise readme of op-cal-tflops Created-by: wuweiqiang24 Commit-by: wuweiqiang24 Merged-by: ascend-robot Description: add usage constraint for op-cal-tflops See merge request: Ascend/MindSpeed!2870	11 个月前
pipeline-parallel.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
preprocess-eod-data.md	Add offline pad_data Co-authored-by: wuweiqiang24<wuweiqiang11@huawei.com> # message auto-generated for no-merge-commit merge: !2938 merge revise_preprocess_data into master Add offline pad_data Created-by: wuweiqiang24 Commit-by: wuweiqiang24 Merged-by: ascend-robot Description: 增加离线预处理pack数据集功能，可提前将数据padding到2\*CP倍，在线使用CP功能时可节约padding部分耗时 * 精度与非离线padding版本存在一定差异  * Llama2-7b，单机16k，GBS=8场景下，性能提升4.8%  See merge request: Ascend/MindSpeed!2938	9 个月前
recomputation.md	[Bugfix][Docs]recomputation Co-authored-by: passioningsun<heqianwei@h-partners.com> # message auto-generated for no-merge-commit merge: !2943 merge master into master [Bugfix][Docs]recomputation Created-by: passioningsun Commit-by: passioningsun Merged-by: ascend-robot Description: [BugFix]Readme&torch_extension&swap-optimizer&swap_attention [BugFix]Readme&torch_extension&swap-optimizer&swap_attention Merge branch 'master' of https://gitcode.com/passioningsun/MindSpeed [Bugfix][Docs]recomputation See merge request: Ascend/MindSpeed!2943	10 个月前
recompute_independent_pipelining.md	【问题单】补充特性限制说明 Co-authored-by: libaokui<libaokui@huawei.com> # message auto-generated for no-merge-commit merge: !2998 merge wtd into master 【问题单】补充特性限制说明 Created-by: libaokui Commit-by: libaokui Merged-by: ascend-robot Description: 1. 暂不支持Megatron原生特性defer-embedding-wgrad-compute 2. recompute-in-bubble不支持与swap-attention、no-overlap-p2p-communication、overlap-grad-reducer不兼容 3. recompute-in-advance不支持no-overlap-p2p-communication、overlap-grad-reduce不兼容 详见以下报告： https://wiki.huawei.com/domains/73366/wiki/264623/WIKI202511139013241?title=4----overlap-grad-reduce---overlap-param-gather See merge request: Ascend/MindSpeed!2998	9 个月前
reset-bucket-group-order.md	重排bucket group；在模型定义顺序与执行顺序混乱的情况下，可以明显提升开启overlap param gather后的吞吐 Co-authored-by: zjuCC<caizeyong1@huawei.com> # message auto-generated for no-merge-commit merge: merge feature/reset_bucket_group into master 重排bucket group；在模型定义顺序与执行顺序混乱的情况下，可以明显提升开启overlap param gather后的吞吐 Created-by: zjuCC Commit-by: zjuCC Merged-by: ascend-robot Description: 在大模型的训练过程中，模型的定义顺序和执行不一致是非常常见的问题，尤其是重定义常见的transformer组件或者使用多模态大模型时。这直接导致overlap-param-gather参数时会出现精度问题和计算通信串行的问题。目前megatron 0.12.1的方案解决了精度问题，然而不可避免地会出现计算和通信串行的问题。 具体实验设计参考https://clouddocs.huawei.com/wapp/doc/a5f778cd-fa07-49fc-9b23-3d14c4c67c85 See merge request: Ascend/MindSpeed!2840	11 个月前
reuse-fp32-param.md	docs: 添加reuse-fp32-param使用约束，断点续训不支持改切分及卡数 Co-authored-by: lizhi<lizhi166@huawei.com> # message auto-generated for no-merge-commit merge: !3101 merge bug-master into master docs: 添加reuse-fp32-param使用约束，断点续训不支持改切分及卡数 Created-by: liz_hw Commit-by: lizhi Merged-by: ascend-robot Description: docs:reuse-fp32-param特性，断点续训扩大dp前反向精度异常 See merge request: Ascend/MindSpeed!3101	8 个月前
ring-attention-context-parallel.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
rms_norm.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
rotary-embedding.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
sequence-parallel.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
shared-experts.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
smart_swap.md	feat(smart-swap): simplify the use of smart-swap Co-authored-by: ChenDonYY<caichendong2@huawei.com> # message auto-generated for no-merge-commit merge: !2833 merge master into master feat(smart-swap): simplify the use of smart-swap Created-by: ChenDonYY Commit-by: ChenDonYY Merged-by: ascend-robot Description: fix: simplify the use of smart-swap 1. 实验需要对比，在使能特性前后，Loss精度、吞吐均值、内存占用。2000步Loss精度相对误差要求2%以内。 - Dense模型用例选取：tests_extend/system_tests/feature_tests/coc.sh - 吞吐比对： swap0 recomput1：80.7 swap0 recompute0：87.7 swap1 recomput0：87.0 - 内存比对： swap0 recomput1 [Rank 0] memory (MB) | allocated: 15604.52587890625 | max allocated: 27669.36279296875 | reserved: 30404.0 | max reserved: 30404.0 [Rank 1] memory (MB) | allocated: 15604.52587890625 | max allocated: 27669.36279296875 | reserved: 30404.0 | max reserved: 30404.0 [Rank 4] memory (MB) | allocated: 16116.654296875 | max allocated: 25036.85986328125 | reserved: 26344.0 | max reserved: 26344.0 [Rank 5] memory (MB) | allocated: 16116.654296875 | max allocated: 25036.85986328125 | reserved: 26344.0 | max reserved: 26344.0 swap0 recompute0 [Rank 0] memory (MB) | allocated: 15604.52587890625 | max allocated: 35925.6298828125 | reserved: 37984.0 | max reserved: 37984.0 [Rank 1] memory (MB) | allocated: 15604.52587890625 | max allocated: 35925.6298828125 | reserved: 37984.0 | max reserved: 37984.0 [Rank 4] memory (MB) | allocated: 16116.654296875 | max allocated: 33549.12744140625 | reserved: 35164.0 | max reserved: 35164.0 [Rank 5] memory (MB) | allocated: 16116.654296875 | max allocated: 33549.12744140625 | reserved: 35164.0 | max reserved: 35164.0 swap1 recompute0 [Rank 0] memory (MB) | allocated: 15672.38427734375 | max allocated: 28631.20361328125 | reserved: 36132.0 | max reserved: 36132.0 [Rank 1] memory (MB) | allocated: 15672.38427734375 | max allocated: 28631.20361328125 | reserved: 36132.0 | max reserved: 36132.0 [Rank 4] memory (MB) | allocated: 16188.48046875 | max allocated: 29610.9287109375 | reserved: 33732.0 | max reserved: 33732.0 [Rank 5] memory (MB) | allocated: 16188.48046875 | max allocated: 29610.9287109375 | reserved: 33732.0 | max reserved: 33732.0 - loss比对：  - MOE模型用例选取：tests_extend/system_tests/feature_tests/deepseek_mla.sh - 吞吐比对： swap0：55.2 swap1：56.0 - 内存比对： swap0 [Rank 0] memory (MB) | allocated: 16443.3466796875 | max allocated: 26676.16259765625 | reserved: 32442.0 | max reserved: 32442.0 [Rank 4] memory (MB) | allocated: 25676.61572265625 | max allocated: 36900.34814453125 | reserved: 43500.0 | max reserved: 43500.0 swap1 [Rank 0] memory (MB) | allocated: 16518.9033203125 | max allocated: 27864.86279296875 | reserved: 32240.0 | max reserved: 32240.0 [Rank 4] memory (MB) | allocated: 25781.51123046875 | max allocated: 38881.0888671875 | reserved: 41112.0 | max reserved: 41112.0 - loss比对：  2. 自定义cpp算子（例如atb等）的接入示例。 见docs/features/smart_swap.md。 See merge request: Ascend/MindSpeed!2833	8 个月前
swap-optimizer.md	[BugFix]Readme&torch_extension&swap-optimizer&swap_attention Co-authored-by: passioningsun<heqianwei@h-partners.com> # message auto-generated for no-merge-commit merge: !2923 merge master into master [BugFix]Readme&torch_extension&swap-optimizer&swap_attention Created-by: passioningsun Commit-by: passioningsun Merged-by: ascend-robot Description: [BugFix]Readme&torch_extension&swap-optimizer&swap_attention See merge request: Ascend/MindSpeed!2923	10 个月前
swap_attention.md	[BugFix]Readme&torch_extension&swap-optimizer&swap_attention Co-authored-by: passioningsun<heqianwei@h-partners.com> # message auto-generated for no-merge-commit merge: !2923 merge master into master [BugFix]Readme&torch_extension&swap-optimizer&swap_attention Created-by: passioningsun Commit-by: passioningsun Merged-by: ascend-robot Description: [BugFix]Readme&torch_extension&swap-optimizer&swap_attention See merge request: Ascend/MindSpeed!2923	10 个月前
swiglu.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
te-comparsion-online.md	Support TransformerEngine Co-authored-by: MingzhenWang<wangmingzhen4@huawei.com> Co-authored-by: Muu<koimuu@163.com> Co-authored-by: x30061065<xuyuanhui3@h-partners.com> Co-authored-by: 耿瑞良<gengruiliang@huawei.com> # message auto-generated for no-merge-commit merge: !2947 merge lingqu_master into master Support TransformerEngine Created-by: mingzhenwang Commit-by: mingzhenwang;Muu;MingzhenWang;x30061065;耿瑞良 Merged-by: ascend-robot Description: 1. 支持TELinear层 2. 支持FP8计算，quantmatmul/gmm 3. 支持多种数据类型FP8/HiF8 4. 支持多种量化策略delayed/tensorwise/blockwise/mxfp8 5. TELinear层支持通算融合 See merge request: Ascend/MindSpeed!2947	9 个月前
tensor-parallel-2d.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
tensor-parallel.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
transformer_engine.md	Support TransformerEngine Co-authored-by: MingzhenWang<wangmingzhen4@huawei.com> Co-authored-by: Muu<koimuu@163.com> Co-authored-by: x30061065<xuyuanhui3@h-partners.com> Co-authored-by: 耿瑞良<gengruiliang@huawei.com> # message auto-generated for no-merge-commit merge: !2947 merge lingqu_master into master Support TransformerEngine Created-by: mingzhenwang Commit-by: mingzhenwang;Muu;MingzhenWang;x30061065;耿瑞良 Merged-by: ascend-robot Description: 1. 支持TELinear层 2. 支持FP8计算，quantmatmul/gmm 3. 支持多种数据类型FP8/HiF8 4. 支持多种量化策略delayed/tensorwise/blockwise/mxfp8 5. TELinear层支持通算融合 See merge request: Ascend/MindSpeed!2947	9 个月前
ulysses-context-parallel.md	!2744 fix: fix the non-standard html table Merge pull request !2744 from YE ZHENYUAN/master0814	1 年前
unaligned-pipeline.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
unaligned-ring-context-parallel.md	!2739 feat: verl and unaligned ring readme Merge pull request !2739 from YE ZHENYUAN/master0806	1 年前
unaligned-ulysses-context-parallel.md	!2246 refactor: context parallel refactor Merge pull request !2246 from YE ZHENYUAN/master	1 年前
unaligned_linear.md	!2593 Readme update！ Merge pull request !2593 from yangjie/master	1 年前
variable_seq_lengths.md	!2431 【feat】Zero_Memory update. Merge pull request !2431 from yangjie/master	1 年前
verl-ckpt-acceleration.md	perf(verl ckpt): ckpt load and save acceleration Co-authored-by: 李鸣沼<lmztju@126.com> # message auto-generated for no-merge-commit merge: !3074 merge verl_load_and_save_ckpt into master perf(verl ckpt): ckpt load and save acceleration Created-by: lmztju Commit-by: 李鸣沼;l30057177 Merged-by: ascend-robot Description: **测试场景1：** verl+megatron后端dapo-qwen3-30b 910A3双机加载和保存ckpt 本地存储  See merge request: Ascend/MindSpeed!3074	8 个月前
verl-ckpt-acceleration.png	perf(verl ckpt): ckpt load and save acceleration Co-authored-by: 李鸣沼<lmztju@126.com> # message auto-generated for no-merge-commit merge: !3074 merge verl_load_and_save_ckpt into master perf(verl ckpt): ckpt load and save acceleration Created-by: lmztju Commit-by: 李鸣沼;l30057177 Merged-by: ascend-robot Description: **测试场景1：** verl+megatron后端dapo-qwen3-30b 910A3双机加载和保存ckpt 本地存储  See merge request: Ascend/MindSpeed!3074	8 个月前
virtual-optimizer.md	!2814 docs: swap and virtual optimizer Merge pull request !2814 from 邓佳/core_r0.12.1_docs	11 个月前
virtual-pipeline-parallel.md	[docs] Modify the Verl documentation Co-authored-by: wangjinyi6<wangjinyi6@huawei.com> # message auto-generated for no-merge-commit merge: !3168 merge master into master [docs] Modify the Verl documentation Created-by: wangjinyi6 Commit-by: wangjinyi6 Merged-by: ascend-robot Description: 修改 verl 文档，添加mbridge使用说明 See merge request: Ascend/MindSpeed!3168	7 个月前