0
代码
代码
Issues
Pull Requests
讨论
Wiki
分析
0
ascend-robotascend-robotdocs(pytorch): update feature doc layerwise_disaggregated_training
ff844edb创建于 13 天前历史提交
文件最后提交记录最后更新时间
layerwise_disaggregated_training_stage.png
docs(pytorch): add feature doc layerwise_disaggregated_training.md Co-authored-by: xuguoliang3<xuguoliang3@huawei.com> # message auto-generated for no-merge-commit merge: !4271 merge 20260302_feature_doc into master docs(pytorch): add feature doc layerwise_disaggregated_training.md Created-by: xuguoliang3 Commit-by: xuguoliang3 Merged-by: ascend-robot Description: # What this PR does / why we need it? The new features of edge-cloud layerwise disaggregated training are added to this document as well as operation instructions. # Does this PR introduce any user-facing change? Documentation paths have changed: Added docs/zh/pytorch/features/mcore/layerwise_disaggregated_training.md Added docs/zh/pytorch/figures/ldt_sft/layerwise_disaggregated_training_stage.png Added docs/zh/pytorch/figures/ldt_sft/pipeline_chart.png Added docs/zh/pytorch/training/finetune/mcore/layerwise_disaggregated_training.md # How was this patch tested? This PR only involves documentation modifications and does not require test cases. See merge request: Ascend/MindSpeed-LLM!42712 个月前
ldt_dp.png
docs(pytorch): update feature doc layerwise_disaggregated_training Co-authored-by: xuguoliang3<xuguoliang3@huawei.com> # message auto-generated for no-merge-commit merge: !4453 merge 20260507_iter5_doc into master docs(pytorch): update feature doc layerwise_disaggregated_training Created-by: xuguoliang3 Commit-by: xuguoliang3 Merged-by: ascend-robot Description: ## What this PR does / why we need it? update feature doc layerwise_disaggregated_training, support different DP or DP&TP configuration on edge and cloud ## Does this PR introduce any user-facing change? Please describe whether the PR will result in any user-facing usage changes. If there is related documentation, please specify its path. ## How was this patch tested? Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations. See merge request: Ascend/MindSpeed-LLM!445313 天前
ldt_tp.png
docs(pytorch): update docs for layerwise_disaggregated_training Co-authored-by: yanzhenghao<yanzhenghao2@huawei.com> Co-authored-by: f00620112<fangminghao@huawei.com> # message auto-generated for no-merge-commit merge: !4388 merge 20260408_iter3_feature_doc into master docs(pytorch): update docs for layerwise_disaggregated_training Created-by: xuguoliang3 Commit-by: f00620112;yanzhenghao;xuguoliang3 Merged-by: ascend-robot Description: ## What this PR does / why we need it? update docs for layerwise_disaggregated_training feature doc and guide, support different TP configuration for the edge and cloud. ## Does this PR introduce any user-facing change? update docs/zh/pytorch/features/mcore/layerwise_disaggregated_training.md update docs/zh/pytorch/training/finetune/mcore/layerwise_disaggregated_training.md ## How was this patch tested? Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations. See merge request: Ascend/MindSpeed-LLM!43881 个月前
ldt_vdp.png
docs(pytorch): update feature doc layerwise_disaggregated_training Co-authored-by: xuguoliang3<xuguoliang3@huawei.com> # message auto-generated for no-merge-commit merge: !4453 merge 20260507_iter5_doc into master docs(pytorch): update feature doc layerwise_disaggregated_training Created-by: xuguoliang3 Commit-by: xuguoliang3 Merged-by: ascend-robot Description: ## What this PR does / why we need it? update feature doc layerwise_disaggregated_training, support different DP or DP&TP configuration on edge and cloud ## Does this PR introduce any user-facing change? Please describe whether the PR will result in any user-facing usage changes. If there is related documentation, please specify its path. ## How was this patch tested? Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations. See merge request: Ascend/MindSpeed-LLM!445313 天前
ldt_vdp_gen_ranks.png
docs(pytorch): update feature doc layerwise_disaggregated_training Co-authored-by: xuguoliang3<xuguoliang3@huawei.com> # message auto-generated for no-merge-commit merge: !4453 merge 20260507_iter5_doc into master docs(pytorch): update feature doc layerwise_disaggregated_training Created-by: xuguoliang3 Commit-by: xuguoliang3 Merged-by: ascend-robot Description: ## What this PR does / why we need it? update feature doc layerwise_disaggregated_training, support different DP or DP&TP configuration on edge and cloud ## Does this PR introduce any user-facing change? Please describe whether the PR will result in any user-facing usage changes. If there is related documentation, please specify its path. ## How was this patch tested? Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations. See merge request: Ascend/MindSpeed-LLM!445313 天前
ldt_vtp.png
docs(pytorch): update docs for layerwise_disaggregated_training Co-authored-by: yanzhenghao<yanzhenghao2@huawei.com> Co-authored-by: f00620112<fangminghao@huawei.com> # message auto-generated for no-merge-commit merge: !4388 merge 20260408_iter3_feature_doc into master docs(pytorch): update docs for layerwise_disaggregated_training Created-by: xuguoliang3 Commit-by: f00620112;yanzhenghao;xuguoliang3 Merged-by: ascend-robot Description: ## What this PR does / why we need it? update docs for layerwise_disaggregated_training feature doc and guide, support different TP configuration for the edge and cloud. ## Does this PR introduce any user-facing change? update docs/zh/pytorch/features/mcore/layerwise_disaggregated_training.md update docs/zh/pytorch/training/finetune/mcore/layerwise_disaggregated_training.md ## How was this patch tested? Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations. See merge request: Ascend/MindSpeed-LLM!43881 个月前
pipeline_chart.png
docs(pytorch): add feature doc layerwise_disaggregated_training.md Co-authored-by: xuguoliang3<xuguoliang3@huawei.com> # message auto-generated for no-merge-commit merge: !4271 merge 20260302_feature_doc into master docs(pytorch): add feature doc layerwise_disaggregated_training.md Created-by: xuguoliang3 Commit-by: xuguoliang3 Merged-by: ascend-robot Description: # What this PR does / why we need it? The new features of edge-cloud layerwise disaggregated training are added to this document as well as operation instructions. # Does this PR introduce any user-facing change? Documentation paths have changed: Added docs/zh/pytorch/features/mcore/layerwise_disaggregated_training.md Added docs/zh/pytorch/figures/ldt_sft/layerwise_disaggregated_training_stage.png Added docs/zh/pytorch/figures/ldt_sft/pipeline_chart.png Added docs/zh/pytorch/training/finetune/mcore/layerwise_disaggregated_training.md # How was this patch tested? This PR only involves documentation modifications and does not require test cases. See merge request: Ascend/MindSpeed-LLM!42712 个月前