| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
[Bugfix]MM_FSDP2:MM_FSDP2:cu_seq_len needs calculating before position_ids split when using Ulysses Co-authored-by: zs-derrick1<1434012475@qq.com> # message auto-generated for no-merge-commit merge: !2161 merge fsdp2_dev3 into fsdp2_dev [Bugfix]MM_FSDP2:MM_FSDP2:cu_seq_len needs calculating before position_ids split when using Ulysses Created-by: zs-derrick1 Commit-by: zs-derrick1 Merged-by: ascend-robot Description: ## Motivation MM_FSDP2:cu_seq_len needs calculating before position_ids split when using Ulysses ## Modification cu_seq_len needs calculating before position_ids split when using Ulysses ## 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!2161 | 4 个月前 | |
[Feature] qwen3.5 support expert parallel Co-authored-by: young256<liumingyang16@huawei.com> # message auto-generated for no-merge-commit merge: !2210 merge fsdp2_dev_qwen35_publish into fsdp2_dev [Feature] qwen3.5 support expert parallel Created-by: young256 Commit-by: young256 Merged-by: ascend-robot Description: ## Motivation Qwen3.5 MoE requires enabling expert parallelism at the 397B parameter scale to prevent Out‑of‑Memory (OOM) errors. ## Modification Supporting expert parallel for Qwen3.5 MoE ## Self-test (Optional) not involve yet ## BC-breaking (Optional) not involve ## 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!2210 | 4 个月前 | |
【bugfix】fix copy bug Co-authored-by: weixin_44031810<gaojie75@huawei.com> # message auto-generated for no-merge-commit merge: !2148 merge fsdp2_dev into fsdp2_dev 【bugfix】fix copy bug Created-by: gaojie_ Commit-by: weixin_44031810 Merged-by: ascend-robot Description: ## Motivation correct copy error ## Modification Please briefly describe what modification is made in this PR. ## 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!2148 | 4 个月前 | |
[Feature] MM_FSDP2: Add attention patch for transformers. And add ring & usp for qwen3vl Co-authored-by: zs-derrick1<1434012475@qq.com> # message auto-generated for no-merge-commit merge: !2120 merge fsdp2_dev2 into fsdp2_dev [Feature] MM_FSDP2: Add attention patch for transformers. And add ring & usp for qwen3vl Created-by: zs-derrick1 Commit-by: zs-derrick1 Merged-by: ascend-robot Description: ## Motivation Add attention patch for transformers. And add ring & usp for qwen3vl ## Modification 1.Add transformers attention patch to support cp(ulysses/ring/usp) 2.add ring&usp for qwen3vl ## 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!2120 | 4 个月前 | |
[feature] recompute support use_reentrant Co-authored-by: weixin_44031810<gaojie75@huawei.com> # message auto-generated for no-merge-commit merge: !2216 merge fsdp2_dev into fsdp2_dev [feature] recompute support use_reentrant Created-by: gaojie_ Commit-by: weixin_44031810 Merged-by: ascend-robot Description: ## Motivation Please describe the motivation of this PR and the goal you want to achieve through this PR. ## Modification Please briefly describe what modification is made in this PR. ## 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!2216 | 4 个月前 |
| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
| 4 个月前 | ||
| 4 个月前 | ||
| 4 个月前 | ||
| 4 个月前 | ||
| 4 个月前 |