0
代码
代码
Issues
Pull Requests
讨论
Wiki
分析
0
ascend-robotascend-robot[bugfix] fix attention mask type for wan
3b9b9639创建于 2月26日历史提交
文件最后提交记录最后更新时间
data.json
[Feature] add truncate_t_by_sp to data config Co-authored-by: 林明哲<linmingzhe3@huawei.com> # message auto-generated for no-merge-commit merge: !1612 merge 1015fixcp into master [Feature] add truncate_t_by_sp to data config Created-by: LinMingZhe Commit-by: 林明哲 Merged-by: ascend-robot Description: ## Motivation add truncate_t_by_sp to data config ## Modification add truncate_t_by_sp to data config ## Checklist **Before PR**: - [ ] The new code needs to comply with the Clean Code specification. - [ ] The PR content is self-checked, and the expression can be clear and the writing standardized **After PR**: - [ ] CLA has been signed and all committers have signed the CLA in this PR. - [ ] The ci-pipeline is passed, Code Check is passed. See merge request: Ascend/MindSpeed-MM!16127 个月前
inference.sh
[bugfix] fix attention mask type for wan Co-authored-by: js1234567<jiangshuo9@h-partners.com> # message auto-generated for no-merge-commit merge: !2211 merge master into master [bugfix] fix attention mask type for wan Created-by: js1234567 Commit-by: js1234567 Merged-by: ascend-robot Description: ## Motivation 开启CP得情况下，Mindspeed默认支持得attention mask type为casual，导致推理结果为花屏。 ## Modification Readme 推理shell脚本 ## Self-test (Optional) If modifications to this PR may cause/fix function/accuracy/performance DTSs/issues, a self-inspection record needs to be attached. ## BC-breaking (Optional) If there are compatibility issues, such as dependencies on cann/torch_npu versions, they need to be explained in the PR. ## Checklist **Before PR**: - [x] The new code needs to comply with the Clean Code specification. - [x] The PR content is self-checked, and the expression can be clear and the writing standardized **After PR**: - [ ] CLA has been signed and all committers have signed the CLA in this PR. - [ ] The ci-pipeline is passed, Code Check is passed. See merge request: Ascend/MindSpeed-MM!22113 个月前
inference_model.json
[Feature] wan2.2 ti2v 5B shell and config Co-authored-by: wumengjie (D)<wumengjie6@huawei.com> # message auto-generated for no-merge-commit merge: merge wan_22_config into master [Feature] wan2.2 ti2v 5B shell and config Created-by: meng-coding Commit-by: wumengjie (D) Merged-by: ascend-robot Description: ## Motivation wan2.2 ti2v 5B shell and config ## Modification wan2.2 ti2v 5B shell and config ## Self-test (Optional) If modifications to this PR may cause/fix function/accuracy/performance DTSs/issues, a self-inspection record needs to be attached. ## BC-breaking (Optional) If there are compatibility issues, such as dependencies on cann/torch_npu versions, they need to be explained in the PR. ## Checklist **Before PR**: - [ ] The new code needs to comply with the Clean Code specification. - [ ] The PR content is self-checked, and the expression can be clear and the writing standardized **After PR**: - [ ] CLA has been signed and all committers have signed the CLA in this PR. - [ ] The ci-pipeline is passed, Code Check is passed. See merge request: Ascend/MindSpeed-MM!15478 个月前
pretrain.sh
[bugfix] fix attention mask type for wan Co-authored-by: js1234567<jiangshuo9@h-partners.com> # message auto-generated for no-merge-commit merge: !2211 merge master into master [bugfix] fix attention mask type for wan Created-by: js1234567 Commit-by: js1234567 Merged-by: ascend-robot Description: ## Motivation 开启CP得情况下，Mindspeed默认支持得attention mask type为casual，导致推理结果为花屏。 ## Modification Readme 推理shell脚本 ## Self-test (Optional) If modifications to this PR may cause/fix function/accuracy/performance DTSs/issues, a self-inspection record needs to be attached. ## BC-breaking (Optional) If there are compatibility issues, such as dependencies on cann/torch_npu versions, they need to be explained in the PR. ## Checklist **Before PR**: - [x] The new code needs to comply with the Clean Code specification. - [x] The PR content is self-checked, and the expression can be clear and the writing standardized **After PR**: - [ ] CLA has been signed and all committers have signed the CLA in this PR. - [ ] The ci-pipeline is passed, Code Check is passed. See merge request: Ascend/MindSpeed-MM!22113 个月前
pretrain_model.json
[Feature] Add the muon optimizer and adapt it for FSDP2. Co-authored-by: hanyyy<hanyue42@huawei.com> # message auto-generated for no-merge-commit merge: !1964 merge master into master [Feature] Add the muon optimizer and adapt it for FSDP2. Created-by: vasileone Commit-by: hanyyy Merged-by: ascend-robot Description: ## Motivation Added the Muon optimizer and enabled its compatibility with FSDP2. ## Modification - Added Muon optimizer implementation with FSDP2 compatibility. - Added patches for Megatron-core to enable zero-code-change injection. - Fixed the issue of Muon optimizer failing under FSDP2 context: sharded parameter handling, gradient sync, and step behavior. ## Self-test (Optional) If modifications to this PR may cause/fix function/accuracy/performance DTSs/issues, a self-inspection record needs to be attached. ## BC-breaking (Optional) If there are compatibility issues, such as dependencies on cann/torch_npu versions, they need to be explained in the PR. ## Checklist **Before PR**: - The new code needs to comply with the Clean Code specification. - The PR content is self-checked, and the expression can be clear and the writing standardized **After PR**: - CLA has been signed and all committers have signed the CLA in this PR. - The ci-pipeline is passed, Code Check is passed. See merge request: Ascend/MindSpeed-MM!19645 个月前