0
代码
代码
Issues
Pull Requests
讨论
Wiki
分析
0
ascend-robotascend-robotcleancode
4bf77ad7创建于 3月30日历史提交
文件最后提交记录最后更新时间
cosyvoice3
[Modify] Enhance performance for TTS model Co-authored-by: AZe_404<wangze62@h-partners.com> # message auto-generated for no-merge-commit merge: !2296 merge tts_perf into master [Modify] Enhance performance for TTS model Created-by: AZe_404 Commit-by: AZe_404 Merged-by: ascend-robot Description: ## What this PR does / why we need it? Enhance performance for TTS model. ## 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-MM!22962 个月前
funasr
[Feature] support FunASR Co-authored-by: js1234567<jiangshuo9@h-partners.com> # message auto-generated for no-merge-commit merge: !2258 merge master into master [Feature] support FunASR Created-by: js1234567 Commit-by: js1234567 Merged-by: ascend-robot Description: ## What this PR does / why we need it? Add FunASR for pretraining in FSDP Support pregathering in FSDP Support DDP with shard size = 1 ## 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-MM!22582 个月前
kimik2_5
[Bugfix] fix kimi k25 moe Co-authored-by: htwang<wanghaitao60@huawei.com> # message auto-generated for no-merge-commit merge: !2314 merge ckpt into master [Bugfix] fix kimi k25 moe Created-by: htwang Commit-by: htwang Merged-by: ascend-robot Description: ## What this PR does / why we need it? fix kimi k25 moe forward when ep = 1 ## 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-MM!23142 个月前
ltx2
[Feature] Add t2v and t2av fine-tuning for LTX 2.0 with FSDP2 backend support Co-authored-by: hanyyy<hanyue42@huawei.com> # message auto-generated for no-merge-commit merge: !2301 merge ltx2_0324 into master [Feature] Add t2v and t2av fine-tuning for LTX 2.0 with FSDP2 backend support Created-by: vasileone Commit-by: hanyyy Merged-by: ascend-robot Description: ## What this PR does / why we need it? 本 PR 在 MindSpeed-MM 的 FSDP2 后端新增 LTX2.0 微调训练能力，提供可直接跑通的 t2v 与 t2av 示例，并补齐视频 / 音视频两套训练 pipeline 模型侧： - 新增 model_id: ltx2，适配 FSDP2 后端 - 新增视频 pipeline - 新增音视频 pipeline ## Modification - mindspeed_mm/fsdp/models/ltx2/ltx2_fsdp2/modeling_ltx2.py：LTX2 的 FSDP2 训练封装 LTX2ForTraining。 - mindspeed_mm/fsdp/data/datasets/ltx2/ltx2_precomputed_dataset.py：ltx2_precomputed fsdp2后端数据集实现。 - examples/fsdp2/ltx2/finetune_ltx2_t2v.sh：t2v torchrun 启动脚本。 - examples/fsdp2/ltx2/finetune_ltx2_t2av.sh：t2av torchrun 启动脚本。 See merge request: Ascend/MindSpeed-MM!23012 个月前
qwen3_5
cleancode Co-authored-by: liyingxuan<liyingxuan3@huawei.com> # message auto-generated for no-merge-commit merge: !2323 merge master into 26.0.0 cleancode Created-by: liyx616 Commit-by: liyingxuan Merged-by: ascend-robot Description: ## What this PR does / why we need it? cleancode整改 ## Does this PR introduce any user-facing change? cleancode整改 ## 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-MM!23232 个月前
qwen3_5_moe
[bugfix]修复RL训练时cp开启alltoall的shape不匹配的问题 Co-authored-by: Miss_min<qiaoxiaomin@huawei.com> # message auto-generated for no-merge-commit merge: !2298 merge master into master [bugfix]修复RL训练时cp开启alltoall的shape不匹配的问题 Created-by: Miss_min Commit-by: Miss_min Merged-by: ascend-robot Description: ## What this PR does / why we need it? Please describe the background and detailed changes of the PR. If it is a bugfix, please attach the related issue. ## 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-MM!22982 个月前
qwen3omni
feat(torch): Initialize Qwen3-Omni model code from transformers library for pure fsdp2 Co-authored-by: liuxi_<liuxi75@huawei.com> # message auto-generated for no-merge-commit merge: !2307 merge qwen3_omni_moe into master feat(torch): Initialize Qwen3-Omni model code from transformers library for pure fsdp2 Created-by: liuxi_ Commit-by: liuxi_ Merged-by: ascend-robot Description: Motivation 初始化Qwen3Omni的模型代码modeling_qwen3_omni_moe.py，本PR中代码来源是transformers库。 Modification 1.修改点：将transformers仓的qwen3omni模型中modeling_qwen3_omni_moe.py导入到文件mindspeed_mm\fsdp\models\qwen3omni\modeling_qwen3_omni_moe.py中；并进行初步裁剪。 2.已验证：构建通过。 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!23072 个月前
qwen3tts
cleancode Co-authored-by: liyingxuan<liyingxuan3@huawei.com> # message auto-generated for no-merge-commit merge: !2323 merge master into 26.0.0 cleancode Created-by: liyx616 Commit-by: liyingxuan Merged-by: ascend-robot Description: ## What this PR does / why we need it? cleancode整改 ## Does this PR introduce any user-facing change? cleancode整改 ## 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-MM!23232 个月前
qwen3vl
Adapt fused operator in Qwen3VL MoE forward Co-authored-by: pangwendong<pangwendong@huawei.com> # message auto-generated for no-merge-commit merge: !2286 merge qwen3vl_fused into master Adapt fused operator in Qwen3VL MoE forward Created-by: WendongPang Commit-by: pangwendong Merged-by: ascend-robot Description: ## What this PR does / why we need it? https://gitcode.com/Ascend/MindSpeed-MM/issues/276 ## 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? Enable deterministic computation (use_deter_comp: true) and train under default configuration and fusion operator respectively. The relative error of loss comparison is less than 10 ^ -5, and the relative error of gradient comparison is less than 10 ^ -4. See merge request: Ascend/MindSpeed-MM!22862 个月前
base_model.py
feat(torch): Squash merge fsdp2_dev into master. Co-authored-by: zs-derrick1<1434012475@qq.com> # message auto-generated for no-merge-commit merge: !2223 merge master into master feat(torch): Squash merge fsdp2_dev into master. Created-by: zs-derrick1 Commit-by: zs-derrick1 Merged-by: ascend-robot Description: ## Motivation Squash merge fsdp2_dev into master. ## Modification fsdp2_dev分支合并至master ## 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**: - [x] CLA has been signed and all committers have signed the CLA in this PR. - [x] The ci-pipeline is passed, Code Check is passed. See merge request: Ascend/MindSpeed-MM!22233 个月前
modelhub.py
feat(torch): Squash merge fsdp2_dev into master. Co-authored-by: zs-derrick1<1434012475@qq.com> # message auto-generated for no-merge-commit merge: !2223 merge master into master feat(torch): Squash merge fsdp2_dev into master. Created-by: zs-derrick1 Commit-by: zs-derrick1 Merged-by: ascend-robot Description: ## Motivation Squash merge fsdp2_dev into master. ## Modification fsdp2_dev分支合并至master ## 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**: - [x] CLA has been signed and all committers have signed the CLA in this PR. - [x] The ci-pipeline is passed, Code Check is passed. See merge request: Ascend/MindSpeed-MM!22233 个月前