ops-nn/norm · wangpengbo26/ops-nn

文件	最后提交记录	最后更新时间
ada_layer_norm	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
ada_layer_norm_grad	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
ada_layer_norm_quant	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
ada_layer_norm_v2	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
add_layer_norm	add cases Co-authored-by: zhaogrug<zhaohai16@h-partners.com> # message auto-generated for no-merge-commit merge: !8026 merge the_rest_of_cases into master add cases Created-by: zhaogrug Commit-by: zhaogrug Merged-by: cann-robot Description: ## 描述新增仓内新版ttk用例 <!--在这里详细描述你的改动，包括改动的原因和所采取的方法。--> ## 关联的Issue [#4386](https://gitcode.com/cann/ops-nn/issues/4386) <!-- 如果这个PR是为了解决特定的Issue，请在这里提供Issue链接。例如：关联Issue #000--> ## 测试冒烟测试，本地验证 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [x] 其他，请描述：用例添加 ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!8026	2 个月前
add_layer_norm_grad	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
add_layer_norm_quant	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
add_layer_norm_quant_v2	docs: fix errors in aclnn API examples Co-authored-by: Chen_HaoWen<chenhaowen12@huawei.com> # message auto-generated for no-merge-commit merge: !7883 merge docs-fix-md-20260723-master into master docs: fix errors in aclnn API examples Created-by: Chen_HaoWen Commit-by: Chen_HaoWen Merged-by: cann-robot Description: ## 描述修正 `master` 分支中 aclnn API 文档及对应 examples 的文本错误，保证文档代码片段与示例源文件一致： - 修正 `aclnnAddLayerNormQuantV2` 日志中的 `aclrtSetDevicefailed` 和 `Deviceto` 拼写问题。 - 将 `aclnnInplaceRenorm` 调用失败日志中的接口名由 `aclnnRenorm` 修正为 `aclnnInplaceRenorm`。 - 修正 `aclnnRmsNormQuant` 示例中的接口名注释，并移除未使用的 `aclScalar` 描述。 - 同步更新 3 个文档文件和 3 个 examples 文件，不涉及算子功能逻辑变更。 ## 关联的Issue Closes #4287 ## 测试 - 已执行 `git diff --check origin/master...HEAD`，检查通过。 - 已定点检查 6 个目标文件，确认旧错误文本不再残留。 - 本次仅修改文档代码片段、日志字符串和注释，未运行算子构建或功能测试。 ## 文档更新 - `norm/add_layer_norm_quant_v2/docs/aclnnAddLayerNormQuantV2.md` - `norm/add_layer_norm_quant_v2/examples/test_aclnn_add_layer_norm_quant_v2.cpp` - `norm/renorm/docs/aclnnRenorm&aclnnInplaceRenorm.md` - `norm/renorm/examples/test_aclnn_inplace_renorm.cpp` - `norm/rms_norm_quant/docs/aclnnRmsNormQuant.md` - `norm/rms_norm_quant/examples/test_aclnn_rms_norm_quant.cpp` ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [x] AI辅助编写 See merge request: cann/ops-nn!7883	3 个月前
add_rms_norm	cleancode: norm 类算子代码去重 Co-authored-by: rk<wangpengbo26@huawei.com> # message auto-generated for no-merge-commit merge: !7977 merge cleancode8 into master cleancode: norm 类算子代码去重 Created-by: wangpengbo26 Commit-by: rk Merged-by: cann-robot Description: ## 描述 ops-nn/norm 系列 arch35 算子代码重复率治理（cleancode）。在严格保持实现逻辑、计算顺序、内存分配顺序完全一致前提下，通过抽取公共实现/分支尾上提/模板统一消除重复，覆盖 4 个算子： - batch_norm_v3 - `batch_norm_v3_regbase_common.h`：`TwoRowAddPartialMeanWithTail` 复用已有 `TwoRowAddPartialMean`（移动定义、去前向声明） - 三个 infer 变体（continuous_a/small_a/small_ab1）的 `Compute` 统一为模板 helper `InferComputeImpl<Self, T>`，经 `self.VFNormalize(...)` 静态派发 - rms_norm_grad：`rms_norm_grad_regbase_dgamma.h` 抽取 `ComputePreDgamma`（EnQue/DeQue/取址/VFCalcPreDgamma/Free），3 处复用 - add_rms_norm：`add_rms_norm_regbase_common.h` 的 `LoadForHandleRemainV1` 三个 if-constexpr 分支公共尾上提 - add_rms_norm_dynamic_quant：`add_rms_norm_dynamic_quant_regbase_common.h` 抽取 VF 单元 `ComputeYAndAbsMaxVF`（归一化输出 + running abs-max），两处复用 ### 等价性说明均为逐操作等价：函数体逐字节搬运，仅作用域/调用形式变化；if constexpr 下每个实例化指令流不变；寄存器声明位置变化不影响操作顺序与内存分配顺序。已用「去空白 md5 比对」确认无标识符级改动。 ## 关联的Issue - https://gitcode.com/cann/ops-nn/issues/4418 ## 测试 - 逐操作等价校验：内联还原 / token 级去空白 md5 比对，确认无标识符改动 - 950 上板编译 + 精度回归：pass ## 文档更新无 ## 类型标签 - [x] 其他，请描述：代码重构（消除重复代码，降低 codecheck 重复率） ## AI/Agent生成声明 - [x] AI辅助编写 See merge request: cann/ops-nn!7977	2 个月前
add_rms_norm_cast	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
add_rms_norm_dynamic_mx_quant	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
add_rms_norm_dynamic_quant	cleancode: norm 类算子代码去重 Co-authored-by: rk<wangpengbo26@huawei.com> # message auto-generated for no-merge-commit merge: !7977 merge cleancode8 into master cleancode: norm 类算子代码去重 Created-by: wangpengbo26 Commit-by: rk Merged-by: cann-robot Description: ## 描述 ops-nn/norm 系列 arch35 算子代码重复率治理（cleancode）。在严格保持实现逻辑、计算顺序、内存分配顺序完全一致前提下，通过抽取公共实现/分支尾上提/模板统一消除重复，覆盖 4 个算子： - batch_norm_v3 - `batch_norm_v3_regbase_common.h`：`TwoRowAddPartialMeanWithTail` 复用已有 `TwoRowAddPartialMean`（移动定义、去前向声明） - 三个 infer 变体（continuous_a/small_a/small_ab1）的 `Compute` 统一为模板 helper `InferComputeImpl<Self, T>`，经 `self.VFNormalize(...)` 静态派发 - rms_norm_grad：`rms_norm_grad_regbase_dgamma.h` 抽取 `ComputePreDgamma`（EnQue/DeQue/取址/VFCalcPreDgamma/Free），3 处复用 - add_rms_norm：`add_rms_norm_regbase_common.h` 的 `LoadForHandleRemainV1` 三个 if-constexpr 分支公共尾上提 - add_rms_norm_dynamic_quant：`add_rms_norm_dynamic_quant_regbase_common.h` 抽取 VF 单元 `ComputeYAndAbsMaxVF`（归一化输出 + running abs-max），两处复用 ### 等价性说明均为逐操作等价：函数体逐字节搬运，仅作用域/调用形式变化；if constexpr 下每个实例化指令流不变；寄存器声明位置变化不影响操作顺序与内存分配顺序。已用「去空白 md5 比对」确认无标识符级改动。 ## 关联的Issue - https://gitcode.com/cann/ops-nn/issues/4418 ## 测试 - 逐操作等价校验：内联还原 / token 级去空白 md5 比对，确认无标识符改动 - 950 上板编译 + 精度回归：pass ## 文档更新无 ## 类型标签 - [x] 其他，请描述：代码重构（消除重复代码，降低 codecheck 重复率） ## AI/Agent生成声明 - [x] AI辅助编写 See merge request: cann/ops-nn!7977	2 个月前
add_rms_norm_dynamic_quant_v2	format cpp Co-authored-by: yang-di52<yangdi52@huawei.com> # message auto-generated for no-merge-commit merge: !6784 merge issue_fix into master format cpp Created-by: yang-di52 Commit-by: yang-di52 Merged-by: cann-robot Description: ## 描述批量刷新cpp代码格式 ## 关联的Issue [#3791](https://gitcode.com/cann/ops-nn/issues/3791) ## 测试 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [x] 其他，请描述：代码格式化 ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!6784	3 个月前
add_rms_norm_quant	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
add_rms_norm_quant_v2	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
batch_norm	fix: comment out dlog_setlevel calls in UT tests for target operators Co-authored-by: liangyuhua<liangyuhua4@huawei.com> # message auto-generated for no-merge-commit merge: !8042 merge fix/comment-out-dlog-setlevel into master fix: comment out dlog_setlevel calls in UT tests for target operators Created-by: liangyuhua Commit-by: liangyuhua Merged-by: cann-robot Description: ## 描述删除 11 个目标算子 UT 测试文件中所有 `dlog_setlevel` 调用（含已注释的），避免 UT 运行时修改全局日志级别影响其他用例，同时清理无用注释代码。修复 `batch_norm_grad_v3` 中变量名 `workspace_size_holer` 拼写错误为 `workspace_size_holder`。涉及的算子及文件： \| 模块 \| 算子 \| 文件 \| 删除数量 \| \|------\|------\|------\|---------\| \| activation \| ConfusionSoftmaxGrad \| test_confusion_softmax_grad_tiling.cpp \| 6 \| \| activation \| LogSoftmaxGrad \| test_log_softmax_grad_tiling.cpp \| 14 \| \| activation \| SoftmaxGrad \| test_softmax_grad_tiling.cpp \| 12 \| \| index \| BroadcastGradientArgs \| test_broadcast_gradient_args_tiling.cpp \| 32 \| \| norm \| BatchNorm \| test_batch_norm_tiling.cpp \| 3 \| \| norm \| BatchNormGrad \| test_batch_norm_grad_tiling.cpp \| 48 \| \| norm \| BatchNormGradV3 \| test_batch_norm_grad_v3_tiling.cpp \| 44 \| \| norm \| LayerNorm \| test_layer_norm_tiling.cpp \| 2 \| \| norm \| LayerNormV3 \| test_layer_norm_v3_tiling.cpp \| 8 \| \| norm \| RmsNormDynamicMxQuant \| test_rms_norm_dynamic_mx_quant_infershape.cpp \| 2 \| \| norm \| SyncBatchNormGatherStats \| test_sync_batch_norm_gather_stats_tiling.cpp \| 12 \| 共计 11 个文件，删除 183 处 `dlog_setlevel` 调用。额外修复： - `batch_norm_grad_v3` 文件中 46 处变量名 `workspace_size_holer` 拼写错误修正为 `workspace_size_holder` - `layer_norm_v3` 文件 license header 从旧格式更新为标准格式注：RmsNormGrad、SoftmaxV2、LogSoftmaxV2 的 `dlog_setlevel` 删除已通过前置 PR 合入 master，本 PR 不再重复修改。 ## 关联的Issue #4396 ## 测试 - [x] pre-commit 全部通过（clang-format + OAT + codespell 等，11 个文件手动扫描确认） - [x] 仅删除 `dlog_setlevel` 调用及修复变量名拼写，不影响测试逻辑 ## 文档更新无需更新文档 ## 类型标签 - [x] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 - [x] AI辅助编写 See merge request: cann/ops-nn!8042	2 个月前
batch_norm_elemt	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
batch_norm_grad	fix: comment out dlog_setlevel calls in UT tests for target operators Co-authored-by: liangyuhua<liangyuhua4@huawei.com> # message auto-generated for no-merge-commit merge: !8042 merge fix/comment-out-dlog-setlevel into master fix: comment out dlog_setlevel calls in UT tests for target operators Created-by: liangyuhua Commit-by: liangyuhua Merged-by: cann-robot Description: ## 描述删除 11 个目标算子 UT 测试文件中所有 `dlog_setlevel` 调用（含已注释的），避免 UT 运行时修改全局日志级别影响其他用例，同时清理无用注释代码。修复 `batch_norm_grad_v3` 中变量名 `workspace_size_holer` 拼写错误为 `workspace_size_holder`。涉及的算子及文件： \| 模块 \| 算子 \| 文件 \| 删除数量 \| \|------\|------\|------\|---------\| \| activation \| ConfusionSoftmaxGrad \| test_confusion_softmax_grad_tiling.cpp \| 6 \| \| activation \| LogSoftmaxGrad \| test_log_softmax_grad_tiling.cpp \| 14 \| \| activation \| SoftmaxGrad \| test_softmax_grad_tiling.cpp \| 12 \| \| index \| BroadcastGradientArgs \| test_broadcast_gradient_args_tiling.cpp \| 32 \| \| norm \| BatchNorm \| test_batch_norm_tiling.cpp \| 3 \| \| norm \| BatchNormGrad \| test_batch_norm_grad_tiling.cpp \| 48 \| \| norm \| BatchNormGradV3 \| test_batch_norm_grad_v3_tiling.cpp \| 44 \| \| norm \| LayerNorm \| test_layer_norm_tiling.cpp \| 2 \| \| norm \| LayerNormV3 \| test_layer_norm_v3_tiling.cpp \| 8 \| \| norm \| RmsNormDynamicMxQuant \| test_rms_norm_dynamic_mx_quant_infershape.cpp \| 2 \| \| norm \| SyncBatchNormGatherStats \| test_sync_batch_norm_gather_stats_tiling.cpp \| 12 \| 共计 11 个文件，删除 183 处 `dlog_setlevel` 调用。额外修复： - `batch_norm_grad_v3` 文件中 46 处变量名 `workspace_size_holer` 拼写错误修正为 `workspace_size_holder` - `layer_norm_v3` 文件 license header 从旧格式更新为标准格式注：RmsNormGrad、SoftmaxV2、LogSoftmaxV2 的 `dlog_setlevel` 删除已通过前置 PR 合入 master，本 PR 不再重复修改。 ## 关联的Issue #4396 ## 测试 - [x] pre-commit 全部通过（clang-format + OAT + codespell 等，11 个文件手动扫描确认） - [x] 仅删除 `dlog_setlevel` 调用及修复变量名拼写，不影响测试逻辑 ## 文档更新无需更新文档 ## 类型标签 - [x] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 - [x] AI辅助编写 See merge request: cann/ops-nn!8042	2 个月前
batch_norm_grad_v3	add cases Co-authored-by: zhaogrug<zhaohai16@h-partners.com> # message auto-generated for no-merge-commit merge: !8026 merge the_rest_of_cases into master add cases Created-by: zhaogrug Commit-by: zhaogrug Merged-by: cann-robot Description: ## 描述新增仓内新版ttk用例 <!--在这里详细描述你的改动，包括改动的原因和所采取的方法。--> ## 关联的Issue [#4386](https://gitcode.com/cann/ops-nn/issues/4386) <!-- 如果这个PR是为了解决特定的Issue，请在这里提供Issue链接。例如：关联Issue #000--> ## 测试冒烟测试，本地验证 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [x] 其他，请描述：用例添加 ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!8026	2 个月前
batch_norm_v3	还原 infer/dgamma 计算的内联写法将 batch_norm_v3 三个 infer 变体的 Compute 与 rms_norm_grad dgamma 计算改回内联实现, 去掉 InferComputeImpl / ComputePreDgamma 两个 helper 间接,使设备侧计算路径更直观; 不改变计算逻辑、计算顺序与内存分配顺序。 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01RDXGVkU6aL6vJjXEPumYHh	2 个月前
bn_training_reduce	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
common	format cpp Co-authored-by: yang-di52<yangdi52@huawei.com> # message auto-generated for no-merge-commit merge: !6784 merge issue_fix into master format cpp Created-by: yang-di52 Commit-by: yang-di52 Merged-by: cann-robot Description: ## 描述批量刷新cpp代码格式 ## 关联的Issue [#3791](https://gitcode.com/cann/ops-nn/issues/3791) ## 测试 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [x] 其他，请描述：代码格式化 ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!6784	3 个月前
deep_norm	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
deep_norm_grad	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
dua_quantize_add_layer_norm	format cpp Co-authored-by: yang-di52<yangdi52@huawei.com> # message auto-generated for no-merge-commit merge: !6784 merge issue_fix into master format cpp Created-by: yang-di52 Commit-by: yang-di52 Merged-by: cann-robot Description: ## 描述批量刷新cpp代码格式 ## 关联的Issue [#3791](https://gitcode.com/cann/ops-nn/issues/3791) ## 测试 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [x] 其他，请描述：代码格式化 ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!6784	3 个月前
euclidean_norm	TF plugin 算子迁移 Co-authored-by: sunchun<sunchun4@h-partners.com> # message auto-generated for no-merge-commit merge: !7619 merge nn3 into master TF plugin 算子迁移 Created-by: sunchun Commit-by: sunchun Merged-by: cann-robot Description: ## 描述 TF plugin 迁移 ## 关联的Issue https://gitcode.com/cann/ops-nn/issues/4196 ## 测试 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [x] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7619	3 个月前
gemma_rms_norm	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
group_norm	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
group_norm_grad	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
group_norm_silu	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
group_norm_silu_quant	non_zero_with_value, multilabel_margin_loss和group_norm_silu_quant下一代算子支持 Co-authored-by: tangpingchuan<tangpingchuan@huawei.com> # message auto-generated for no-merge-commit merge: !7599 merge master into master non_zero_with_value, multilabel_margin_loss和group_norm_silu_quant下一代算子支持 Created-by: zl_hw Commit-by: tangpingchuan Merged-by: cann-robot Description: ## 描述支持下一代non_zero_with_value, multilabel_margin_loss和group_norm_silu_quant算子 ## 关联的Issue https://gitcode.com/cann/ops-nn/issues/4221 ## 测试 ut/st ## 文档更新 README.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [x] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [x] AI辅助编写 See merge request: cann/ops-nn!7599	3 个月前
group_norm_swish	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
group_norm_swish_grad	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
group_norm_v2	add_cases Co-authored-by: zhaogrug<zhaohai16@h-partners.com> # message auto-generated for no-merge-commit merge: !7583 merge add_cases into master add_cases Created-by: zhaogrug Commit-by: zhaogrug Merged-by: cann-robot Description: ## 描述新增仓内新版ttk用例 <!--在这里详细描述你的改动，包括改动的原因和所采取的方法。--> ## 关联的Issue [#4223](https://gitcode.com/cann/ops-nn/issues/4223) <!-- 如果这个PR是为了解决特定的Issue，请在这里提供Issue链接。例如：关联Issue #000--> ## 测试冒烟测试，本地验证 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [x] 其他，请描述：用例添加 ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7583	3 个月前
in_infer_v2	fix cleancode Co-authored-by: alfengyuan<yyfgdut@gmail.com> # message auto-generated for no-merge-commit merge: !8050 merge master into master fix cleancode Created-by: alfengyuan Commit-by: alfengyuan Merged-by: cann-robot Description: ## 描述 <!--在这里详细描述你的改动，包括改动的原因和所采取的方法。--> fix cleancode ## 关联的Issue <!-- 如果这个PR是为了解决特定的Issue，请在这里提供Issue链接。例如：关联Issue #000--> cleancode修改，不涉及功能，不涉及文档 ## 测试 <!--描述进行了哪些测试来验证你的改动。--> cleancode修改，不涉及功能，不涉及文档 ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> cleancode修改，不涉及功能，不涉及文档 ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [x] 其他，请描述：纯cleancode ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!8050	2 个月前
in_training_update_grad	norm: add arch35(Ascend950) InTrainingUpdateGrad/InstanceNormGrad/L2NormalizeGrad Co-authored-by: tangpingchuan<tangpingchuan@huawei.com> # message auto-generated for no-merge-commit merge: !7965 merge feat/norm-grad-ops-arch35 into master norm: add arch35(Ascend950) InTrainingUpdateGrad/InstanceNormGrad/L2NormalizeGrad Created-by: zl_hw Commit-by: tangpingchuan Merged-by: cann-robot Description: ## 描述 InTrainingUpdateGrad/InstanceNormGrad/L2NormalizeGrad算子下一代支持 ## 关联的Issue https://gitcode.com/cann/ops-nn/issues/4371 ## 测试 ut/st/冒烟 ## 文档更新 README ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [x] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [x] AI辅助编写 See merge request: cann/ops-nn!7965	2 个月前
inplace_add_layer_norm	modified md files(for readability improvement) Co-authored-by: duhuiping<duhuiping@h-partners.com> # message auto-generated for no-merge-commit merge: !7592 merge master into master modified md files(for readability improvement) Created-by: gitee-duhuiping Commit-by: duhuiping Merged-by: cann-robot Description: ## 描述 Optimized the function description and parameter description. Optimized the calculation formula. Modified the description of interface invoking, deleted repeated content, and corrected parameter names. ## 关联的Issue [#4236](https://gitcode.com/cann/ops-nn/issues/4236) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 md files ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7592	3 个月前
inplace_add_rms_norm	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
instance_norm	add cases Co-authored-by: zhaogrug<zhaohai16@h-partners.com> # message auto-generated for no-merge-commit merge: !8026 merge the_rest_of_cases into master add cases Created-by: zhaogrug Commit-by: zhaogrug Merged-by: cann-robot Description: ## 描述新增仓内新版ttk用例 <!--在这里详细描述你的改动，包括改动的原因和所采取的方法。--> ## 关联的Issue [#4386](https://gitcode.com/cann/ops-nn/issues/4386) <!-- 如果这个PR是为了解决特定的Issue，请在这里提供Issue链接。例如：关联Issue #000--> ## 测试冒烟测试，本地验证 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [x] 其他，请描述：用例添加 ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!8026	2 个月前
instance_norm_grad	norm: add arch35(Ascend950) InTrainingUpdateGrad/InstanceNormGrad/L2NormalizeGrad Co-authored-by: tangpingchuan<tangpingchuan@huawei.com> # message auto-generated for no-merge-commit merge: !7965 merge feat/norm-grad-ops-arch35 into master norm: add arch35(Ascend950) InTrainingUpdateGrad/InstanceNormGrad/L2NormalizeGrad Created-by: zl_hw Commit-by: tangpingchuan Merged-by: cann-robot Description: ## 描述 InTrainingUpdateGrad/InstanceNormGrad/L2NormalizeGrad算子下一代支持 ## 关联的Issue https://gitcode.com/cann/ops-nn/issues/4371 ## 测试 ut/st/冒烟 ## 文档更新 README ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [x] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [x] AI辅助编写 See merge request: cann/ops-nn!7965	2 个月前
instance_norm_v3	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
l2_normalize_grad	norm: add arch35(Ascend950) InTrainingUpdateGrad/InstanceNormGrad/L2NormalizeGrad Co-authored-by: tangpingchuan<tangpingchuan@huawei.com> # message auto-generated for no-merge-commit merge: !7965 merge feat/norm-grad-ops-arch35 into master norm: add arch35(Ascend950) InTrainingUpdateGrad/InstanceNormGrad/L2NormalizeGrad Created-by: zl_hw Commit-by: tangpingchuan Merged-by: cann-robot Description: ## 描述 InTrainingUpdateGrad/InstanceNormGrad/L2NormalizeGrad算子下一代支持 ## 关联的Issue https://gitcode.com/cann/ops-nn/issues/4371 ## 测试 ut/st/冒烟 ## 文档更新 README ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [x] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [x] AI辅助编写 See merge request: cann/ops-nn!7965	2 个月前
layer_norm	fix: comment out dlog_setlevel calls in UT tests for target operators Co-authored-by: liangyuhua<liangyuhua4@huawei.com> # message auto-generated for no-merge-commit merge: !8042 merge fix/comment-out-dlog-setlevel into master fix: comment out dlog_setlevel calls in UT tests for target operators Created-by: liangyuhua Commit-by: liangyuhua Merged-by: cann-robot Description: ## 描述删除 11 个目标算子 UT 测试文件中所有 `dlog_setlevel` 调用（含已注释的），避免 UT 运行时修改全局日志级别影响其他用例，同时清理无用注释代码。修复 `batch_norm_grad_v3` 中变量名 `workspace_size_holer` 拼写错误为 `workspace_size_holder`。涉及的算子及文件： \| 模块 \| 算子 \| 文件 \| 删除数量 \| \|------\|------\|------\|---------\| \| activation \| ConfusionSoftmaxGrad \| test_confusion_softmax_grad_tiling.cpp \| 6 \| \| activation \| LogSoftmaxGrad \| test_log_softmax_grad_tiling.cpp \| 14 \| \| activation \| SoftmaxGrad \| test_softmax_grad_tiling.cpp \| 12 \| \| index \| BroadcastGradientArgs \| test_broadcast_gradient_args_tiling.cpp \| 32 \| \| norm \| BatchNorm \| test_batch_norm_tiling.cpp \| 3 \| \| norm \| BatchNormGrad \| test_batch_norm_grad_tiling.cpp \| 48 \| \| norm \| BatchNormGradV3 \| test_batch_norm_grad_v3_tiling.cpp \| 44 \| \| norm \| LayerNorm \| test_layer_norm_tiling.cpp \| 2 \| \| norm \| LayerNormV3 \| test_layer_norm_v3_tiling.cpp \| 8 \| \| norm \| RmsNormDynamicMxQuant \| test_rms_norm_dynamic_mx_quant_infershape.cpp \| 2 \| \| norm \| SyncBatchNormGatherStats \| test_sync_batch_norm_gather_stats_tiling.cpp \| 12 \| 共计 11 个文件，删除 183 处 `dlog_setlevel` 调用。额外修复： - `batch_norm_grad_v3` 文件中 46 处变量名 `workspace_size_holer` 拼写错误修正为 `workspace_size_holder` - `layer_norm_v3` 文件 license header 从旧格式更新为标准格式注：RmsNormGrad、SoftmaxV2、LogSoftmaxV2 的 `dlog_setlevel` 删除已通过前置 PR 合入 master，本 PR 不再重复修改。 ## 关联的Issue #4396 ## 测试 - [x] pre-commit 全部通过（clang-format + OAT + codespell 等，11 个文件手动扫描确认） - [x] 仅删除 `dlog_setlevel` 调用及修复变量名拼写，不影响测试逻辑 ## 文档更新无需更新文档 ## 类型标签 - [x] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 - [x] AI辅助编写 See merge request: cann/ops-nn!8042	2 个月前
layer_norm_grad	TF plugin 迁移 Co-authored-by: sunchun<sunchun4@h-partners.com> # message auto-generated for no-merge-commit merge: !7632 merge nn4 into master TF plugin 迁移 Created-by: sunchun Commit-by: sunchun Merged-by: cann-robot Description: ## 描述 TF plugin 迁移 ## 关联的Issue https://gitcode.com/cann/ops-nn/issues/4196 ## 测试 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [x] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7632	3 个月前
layer_norm_grad_v3	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
layer_norm_quant	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
layer_norm_v3	fix: comment out dlog_setlevel calls in UT tests for target operators Co-authored-by: liangyuhua<liangyuhua4@huawei.com> # message auto-generated for no-merge-commit merge: !8042 merge fix/comment-out-dlog-setlevel into master fix: comment out dlog_setlevel calls in UT tests for target operators Created-by: liangyuhua Commit-by: liangyuhua Merged-by: cann-robot Description: ## 描述删除 11 个目标算子 UT 测试文件中所有 `dlog_setlevel` 调用（含已注释的），避免 UT 运行时修改全局日志级别影响其他用例，同时清理无用注释代码。修复 `batch_norm_grad_v3` 中变量名 `workspace_size_holer` 拼写错误为 `workspace_size_holder`。涉及的算子及文件： \| 模块 \| 算子 \| 文件 \| 删除数量 \| \|------\|------\|------\|---------\| \| activation \| ConfusionSoftmaxGrad \| test_confusion_softmax_grad_tiling.cpp \| 6 \| \| activation \| LogSoftmaxGrad \| test_log_softmax_grad_tiling.cpp \| 14 \| \| activation \| SoftmaxGrad \| test_softmax_grad_tiling.cpp \| 12 \| \| index \| BroadcastGradientArgs \| test_broadcast_gradient_args_tiling.cpp \| 32 \| \| norm \| BatchNorm \| test_batch_norm_tiling.cpp \| 3 \| \| norm \| BatchNormGrad \| test_batch_norm_grad_tiling.cpp \| 48 \| \| norm \| BatchNormGradV3 \| test_batch_norm_grad_v3_tiling.cpp \| 44 \| \| norm \| LayerNorm \| test_layer_norm_tiling.cpp \| 2 \| \| norm \| LayerNormV3 \| test_layer_norm_v3_tiling.cpp \| 8 \| \| norm \| RmsNormDynamicMxQuant \| test_rms_norm_dynamic_mx_quant_infershape.cpp \| 2 \| \| norm \| SyncBatchNormGatherStats \| test_sync_batch_norm_gather_stats_tiling.cpp \| 12 \| 共计 11 个文件，删除 183 处 `dlog_setlevel` 调用。额外修复： - `batch_norm_grad_v3` 文件中 46 处变量名 `workspace_size_holer` 拼写错误修正为 `workspace_size_holder` - `layer_norm_v3` 文件 license header 从旧格式更新为标准格式注：RmsNormGrad、SoftmaxV2、LogSoftmaxV2 的 `dlog_setlevel` 删除已通过前置 PR 合入 master，本 PR 不再重复修改。 ## 关联的Issue #4396 ## 测试 - [x] pre-commit 全部通过（clang-format + OAT + codespell 等，11 个文件手动扫描确认） - [x] 仅删除 `dlog_setlevel` 调用及修复变量名拼写，不影响测试逻辑 ## 文档更新无需更新文档 ## 类型标签 - [x] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 - [x] AI辅助编写 See merge request: cann/ops-nn!8042	2 个月前
layer_norm_v4	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
lp_norm_update	[cannbot] lp_norm_update算子支持Ascend950 Co-authored-by: Davon14272<你çliuwenda4@huawei.com> # message auto-generated for no-merge-commit merge: !7538 merge lpnorm into master [cannbot] lp_norm_update算子支持Ascend950 Created-by: Davon14272 Commit-by: Davon14272 Merged-by: cann-robot Description: ## 描述 A5新增支持lp_norm_update算子，仅涉及GEIR通路 <!--在这里详细描述你的改动，包括改动的原因和所采取的方法。--> ## 关联的Issue <!-- 如果这个PR是为了解决特定的Issue，请在这里提供Issue链接。例如：关联Issue #000--> ## 测试 <!--描述进行了哪些测试来验证你的改动。--> ttk，冒烟 ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [x] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [x] AI辅助编写 See merge request: cann/ops-nn!7538	3 个月前
lp_norm_update_v2	[CANNBot]新增lp_norm_update_v2算子支持Ascend950 Co-authored-by: chen_chenn<chenchen167@huawei.com> # message auto-generated for no-merge-commit merge: !7527 merge LpNromUpdateV2 into master [CANNBot]新增lp_norm_update_v2算子支持Ascend950 Created-by: chen_chenn Commit-by: chen_chenn Merged-by: cann-robot Description: ## 描述新增lp_norm_update_v2算子支持Ascend950 ## 关联的Issue https://gitcode.com/cann/ops-nn/issues/4104 LpNormUpdateV2 是 torch.linalg.vector_norm 的"取根"阶段：接收归约结果 x=Σ\|x\|^p，按 p 选分支（0/1/±inf→恒等，2→sqrt，其他→exp(ln(x)/p)），再与 epsilon 取 max。 ## 测试 TTK 冒烟：23564 冒烟：7738 ## 文档更新 ## 类型标签 - [ ] Bug修复 - [x] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 - [x] AI辅助编写 See merge request: cann/ops-nn!7527	3 个月前
lp_norm_v2	【SoftmaxV2&LogSoftmaxV2&LpNormV2】支持batch一致性 Co-authored-by: zhang-song-rui<zhangsongrui@h-partners.com> # message auto-generated for no-merge-commit merge: !7879 merge feat/batch-invariant-softmax-lpnorm into master 【SoftmaxV2&LogSoftmaxV2&LpNormV2】支持batch一致性 Created-by: zhang-song-rui Commit-by: zhang-song-rui Merged-by: cann-robot Description: ## 描述本PR为SoftmaxV2、LogSoftmaxV2和LpNormV2新增batch一致性支持。当确定性级别为3时开启batch一致性。 ### SoftmaxV2 / LogSoftmaxV2 - 在基类中读取确定性级别并设置`enableBatchInvariant_`，同时在DEBUG日志中打印该值。 - batch一致性场景跳过AR小尾轴转置模板，避免转置后归约方式变化导致累加顺序不一致。 - 对合轴后`A0 == 1`的场景，结合reduce轴位置区分“原生AR”和“ARA退化为AR”： - 原生AR仍可选择AR全载或AR重计算模板； - ARA退化场景在batch一致性开启时由AR模板拒绝，继续进入ARA模板，避免切换为AR模板。 - ARA全载和ARA重计算作为低优先级兜底模板，无需额外限制`A0 == 1`；模板选择由优先级更高的AR模板完成区分。 ### LpNormV2 - 根据确定性级别设置Reduce tiling参数`batchInvariant`。 - batch一致性计算由Reduce模板负责，算子kernel计算逻辑无需修改。 - tiling日志打印`batchInvariant`和实际tiling key，便于端到端验证模板选择。 ### opbase依赖本修改依赖opbase的batch一致性Reduce模板能力，`cmake/third_party/opbase.cmake`中的`OPBASE_TAG_ID`已更新为： `text bd20a12ebbd210daccd9f936a9ff4fdd62fa9462` 对应opbase提交：`reduce模版针对batch一致性新增GetInputParam接口`。 ## 关联的Issue 无。 ## 测试 ### SoftmaxV2 Ascend 950 Torch NPU端到端验证所有已执行用例均通过CPU精度检查；完整batch、逐条拆分以及`[3, 5]`非均匀拆分的输出原始字节完全一致。 \| 场景 \| level 0 \| level 3 \| 验证结论 \| \| --- \| --- \| --- \| --- \| \| AR小尾轴，shape=`[8, 8]`，axis=`1` \| `batch=0`，priority=`50`，key=`500` \| `batch=1`，跳过小尾轴模板，priority=`100`，key=`1000` \| batch一致性开启后跳过AR小尾轴，转入AR全载 \| \| 原生AR全载，shape=`[8, 1024]`，axis=`1` \| priority=`100`，key=`1000` \| priority=`100`，key=`1000` \| 原生AR在开关前后保持AR全载 \| \| 原生AR重计算，shape=`[8, 20000]`，axis=`1` \| priority=`200`，key=`2000` \| priority=`200`，key=`2000` \| 原生AR在开关前后保持AR重计算 \| \| ARA退化为AR，shape=`[8, 1024, 1]`，axis=`1` \| priority=`100`，key=`1000` \| priority=`400`，key=`20000` \| level 0退化为AR；level 3保持ARA，并因R方向容量进入ARA重计算 \| \| ARA全载，shape=`[8, 4, 2170]`，axis=`1` \| `batch=0`，priority=`300`，key=`10000` \| `batch=1`，priority=`300`，key=`10000` \| ARA全载模板在开关前后保持不变 \| \| 异dtype且尾轴为1，shape=`[8, 1024, 1]`，FP16输入/FP32输出 \| priority=`100`，key=`1000` \| priority=`400`，key=`20000` \| 异dtype场景下level 3保持ARA通路 \| 日志已确认level 0打印`enableBatchInvariant is: 0`，level 3打印`enableBatchInvariant is: 1`；level 3的小尾轴场景均打印AR small-R disabled日志。 ### LogSoftmaxV2 Ascend 950 Torch NPU端到端验证所有已执行用例均通过CPU精度检查；完整batch、逐条拆分以及`[3, 5]`非均匀拆分的输出原始字节完全一致。 \| 场景 \| level 0 \| level 3 \| 验证结论 \| \| --- \| --- \| --- \| --- \| \| AR小尾轴，shape=`[8, 8]`，axis=`1` \| `batch=0`，priority=`50`，key=`500` \| `batch=1`，跳过小尾轴模板，priority=`100`，key=`1000` \| batch一致性开启后跳过AR小尾轴，转入AR全载 \| \| ARA退化为AR，shape=`[8, 1024, 1]`，axis=`1` \| priority=`100`，key=`1000` \| priority=`400`，key=`20000` \| level 0退化为AR；level 3保持ARA，并进入ARA重计算 \| \| 原生AR重计算，shape=`[8, 20000]`，axis=`1` \| SoftmaxV2同模板对照已验证 \| priority=`200`，key=`2000` \| level 3下原生AR仍保持AR重计算 \| \| ARA全载，shape=`[8, 4, 2170]`，axis=`1` \| `batch=0`，priority=`300`，key=`10000` \| `batch=1`，priority=`300`，key=`10000` \| ARA全载模板在开关前后保持不变 \| \| 异dtype且尾轴为1，shape=`[8, 1024, 1]`，FP16输入/FP32输出 \| priority=`100`，key=`1000` \| priority=`400`，key=`20000` \| 异dtype场景下level 3保持ARA通路 \| 日志已确认level 0打印`enableBatchInvariant is: 0`，level 3打印`enableBatchInvariant is: 1`；level 3的小尾轴场景打印AR small-R disabled日志。 ### LpNormV2 Ascend 950 Torch NPU端到端验证最终矩阵18个用例全部通过，templateNum 0～6均完成level 0/3配对验证。所有level 0日志均显示`batchInvariant: 0`，所有level 3日志均显示`batchInvariant: 1`；同一场景的level 0/3 tiling key均不同。 \| P与模板 \| level 0 key \| level 3 key \| \| --- \| ---: \| ---: \| \| P=0，templateNum=0 \| `10285` \| `12371` \| \| P=1，templateNum=1 \| `67119149` \| `67121235` \| \| P=2，templateNum=2 \| `134228013` \| `134230099` \| \| P=3，templateNum=3 \| `201338961` \| `201338963` \| \| P=-inf，templateNum=4 \| `268445741` \| `268447827` \| \| P=inf，templateNum=5 \| `335554605` \| `335556691` \| \| P=3.5，templateNum=6 \| `402663469` \| `402665555` \| 同时覆盖并通过： - 尾轴为1； - FP16输入、FP32输出； - 非连续输入； - 完整batch、逐条拆分以及`[3, 5]`非均匀拆分的输出原始字节一致； - CPU精度检查； - Reduce pattern与batch bit共同参与tiling key生成。 P=1为1024个绝对值累加，按`1e-3`绝对误差阈值验收，实测最大绝对误差为`0.000396728515625`；其完整batch与两种拆分的输出原始字节完全一致。 ## 文档更新无。 ## 类型标签 - [ ] Bug修复 - [x] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 - [ ] AI辅助编写 See merge request: cann/ops-nn!7879	2 个月前
masked_softmax_with_rel_pos_bias	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
multi_add_rms_norm_dynamic_quant	新增 MSELossV2/PoissonNllLoss/MultiAddRmsNormDynamicQuant算子实现 Co-authored-by: tangpingchuan<tangpingchuan@huawei.com> # message auto-generated for no-merge-commit merge: !7910 merge feat/poisson-nll-loss-arch35 into master 新增 MSELossV2/PoissonNllLoss/MultiAddRmsNormDynamicQuant算子实现 Created-by: zl_hw Commit-by: tangpingchuan Merged-by: cann-robot Description: ## 描述新增 MSELossV2/PoissonNllLoss/MultiAddRmsNormDynamicQuant下一代算子实现 ## 关联的Issue https://gitcode.com/cann/ops-nn/issues/4372 ## 测试 ut/st/冒烟 ## 文档更新 README ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [x] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [x] AI辅助编写 See merge request: cann/ops-nn!7910	2 个月前
norm_common	arch35: DataCopyExtParams迁移及文档错误码修正 Co-authored-by: raoliang_sac<raoliang4@huawei.com> # message auto-generated for no-merge-commit merge: !7482 merge cherry-pick-mr-7319-1784000275331-auto into master arch35: DataCopyExtParams迁移及文档错误码修正 Created-by: raoliang_sac Commit-by: raoliang_sac Merged-by: cann-robot Description: https://gitcode.com/cann/ops-nn/issues/4015 ## 改动内容 ### 1. DataCopyParams → DataCopyExtParams 迁移解决 uint16 在大 UB（如 V121 384KB）下溢出问题，blockLen/srcStride/dstStride 升级为 uint32。涉及 arch35 kernel 文件： - `add_rms_norm_dynamic_mx_quant` arch35 (fp8/fp4 split_r) - `rms_norm_quant_v2` arch35 (regbase_recompute) - `rms_norm_dynamic_mx_quant` arch35 (recompute) 配套适配：`DataCopyPadParams` → `DataCopyPadExtParams<T_X>` ### 2. C-style cast 修复 `add_rms_norm_quant_tiling_arch35.cpp` 中 `(int64_t)` → `static_cast<int64_t>`，符合 G.EXP.14-CPP 规范 ### 3. 文档错误码修正 3 个 rmsnorm+quant 算子文档中 `ACLNN_ERR_INNER_TILING_ERROR(561002)` 改为 `ACLNN_ERR_INNER_NULLPTR(561103)`： - `add_rms_norm_dynamic_mx_quant` - `rms_norm_dynamic_mx_quant` - `rms_norm_quant_v3` ### 4. 修复 GCC 14.2 ICE 参考 https://gitcode.com/cann/community/issues/157 ，将 `castTrait` 定义从模板函数内部 `static constexpr` 提取到命名空间作用域，规避 GCC 14.2 模板实例化时 ICE： - `add_rms_norm_dynamic_mx_quant_common.h` - `rms_norm_dynamic_mx_quant_common.h` ## 涉及文件（10个，全部 arch35） \| 文件 \| 改动内容 \| \|------\|--------\| \| add_rms_norm_dynamic_mx_quant arch35 common.h \| GCC14 ICE 修复 + DataCopyExtParams迁移 \| \| add_rms_norm_dynamic_mx_quant arch35 fp8_split_r.h \| DataCopyExtParams迁移 \| \| add_rms_norm_dynamic_mx_quant arch35 fp4_split_r.h \| DataCopyExtParams迁移 \| \| add_rms_norm_dynamic_mx_quant docs \| 错误码修正 \| \| add_rms_norm_quant tiling_arch35.cpp \| C-style cast修复 \| \| rms_norm_dynamic_mx_quant arch35 common.h \| GCC14 ICE 修复 \| \| rms_norm_dynamic_mx_quant arch35 recompute.h \| DataCopyExtParams迁移 \| \| rms_norm_dynamic_mx_quant docs \| 错误码修正 \| \| rms_norm_quant_v2 arch35 regbase_recompute.h \| DataCopyExtParams迁移 \| \| rms_norm_quant_v3 docs \| 错误码修正 \| See merge request: cann/ops-nn!7482	3 个月前
quantize_add_layer_norm	format cpp Co-authored-by: yang-di52<yangdi52@huawei.com> # message auto-generated for no-merge-commit merge: !6784 merge issue_fix into master format cpp Created-by: yang-di52 Commit-by: yang-di52 Merged-by: cann-robot Description: ## 描述批量刷新cpp代码格式 ## 关联的Issue [#3791](https://gitcode.com/cann/ops-nn/issues/3791) ## 测试 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [x] 其他，请描述：代码格式化 ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!6784	3 个月前
quantized_batch_norm	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
renorm	docs: fix errors in aclnn API examples Co-authored-by: Chen_HaoWen<chenhaowen12@huawei.com> # message auto-generated for no-merge-commit merge: !7883 merge docs-fix-md-20260723-master into master docs: fix errors in aclnn API examples Created-by: Chen_HaoWen Commit-by: Chen_HaoWen Merged-by: cann-robot Description: ## 描述修正 `master` 分支中 aclnn API 文档及对应 examples 的文本错误，保证文档代码片段与示例源文件一致： - 修正 `aclnnAddLayerNormQuantV2` 日志中的 `aclrtSetDevicefailed` 和 `Deviceto` 拼写问题。 - 将 `aclnnInplaceRenorm` 调用失败日志中的接口名由 `aclnnRenorm` 修正为 `aclnnInplaceRenorm`。 - 修正 `aclnnRmsNormQuant` 示例中的接口名注释，并移除未使用的 `aclScalar` 描述。 - 同步更新 3 个文档文件和 3 个 examples 文件，不涉及算子功能逻辑变更。 ## 关联的Issue Closes #4287 ## 测试 - 已执行 `git diff --check origin/master...HEAD`，检查通过。 - 已定点检查 6 个目标文件，确认旧错误文本不再残留。 - 本次仅修改文档代码片段、日志字符串和注释，未运行算子构建或功能测试。 ## 文档更新 - `norm/add_layer_norm_quant_v2/docs/aclnnAddLayerNormQuantV2.md` - `norm/add_layer_norm_quant_v2/examples/test_aclnn_add_layer_norm_quant_v2.cpp` - `norm/renorm/docs/aclnnRenorm&aclnnInplaceRenorm.md` - `norm/renorm/examples/test_aclnn_inplace_renorm.cpp` - `norm/rms_norm_quant/docs/aclnnRmsNormQuant.md` - `norm/rms_norm_quant/examples/test_aclnn_rms_norm_quant.cpp` ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [x] AI辅助编写 See merge request: cann/ops-nn!7883	3 个月前
rms_norm	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
rms_norm_dynamic_mx_quant	fix: comment out dlog_setlevel calls in UT tests for target operators Co-authored-by: liangyuhua<liangyuhua4@huawei.com> # message auto-generated for no-merge-commit merge: !8042 merge fix/comment-out-dlog-setlevel into master fix: comment out dlog_setlevel calls in UT tests for target operators Created-by: liangyuhua Commit-by: liangyuhua Merged-by: cann-robot Description: ## 描述删除 11 个目标算子 UT 测试文件中所有 `dlog_setlevel` 调用（含已注释的），避免 UT 运行时修改全局日志级别影响其他用例，同时清理无用注释代码。修复 `batch_norm_grad_v3` 中变量名 `workspace_size_holer` 拼写错误为 `workspace_size_holder`。涉及的算子及文件： \| 模块 \| 算子 \| 文件 \| 删除数量 \| \|------\|------\|------\|---------\| \| activation \| ConfusionSoftmaxGrad \| test_confusion_softmax_grad_tiling.cpp \| 6 \| \| activation \| LogSoftmaxGrad \| test_log_softmax_grad_tiling.cpp \| 14 \| \| activation \| SoftmaxGrad \| test_softmax_grad_tiling.cpp \| 12 \| \| index \| BroadcastGradientArgs \| test_broadcast_gradient_args_tiling.cpp \| 32 \| \| norm \| BatchNorm \| test_batch_norm_tiling.cpp \| 3 \| \| norm \| BatchNormGrad \| test_batch_norm_grad_tiling.cpp \| 48 \| \| norm \| BatchNormGradV3 \| test_batch_norm_grad_v3_tiling.cpp \| 44 \| \| norm \| LayerNorm \| test_layer_norm_tiling.cpp \| 2 \| \| norm \| LayerNormV3 \| test_layer_norm_v3_tiling.cpp \| 8 \| \| norm \| RmsNormDynamicMxQuant \| test_rms_norm_dynamic_mx_quant_infershape.cpp \| 2 \| \| norm \| SyncBatchNormGatherStats \| test_sync_batch_norm_gather_stats_tiling.cpp \| 12 \| 共计 11 个文件，删除 183 处 `dlog_setlevel` 调用。额外修复： - `batch_norm_grad_v3` 文件中 46 处变量名 `workspace_size_holer` 拼写错误修正为 `workspace_size_holder` - `layer_norm_v3` 文件 license header 从旧格式更新为标准格式注：RmsNormGrad、SoftmaxV2、LogSoftmaxV2 的 `dlog_setlevel` 删除已通过前置 PR 合入 master，本 PR 不再重复修改。 ## 关联的Issue #4396 ## 测试 - [x] pre-commit 全部通过（clang-format + OAT + codespell 等，11 个文件手动扫描确认） - [x] 仅删除 `dlog_setlevel` 调用及修复变量名拼写，不影响测试逻辑 ## 文档更新无需更新文档 ## 类型标签 - [x] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 - [x] AI辅助编写 See merge request: cann/ops-nn!8042	2 个月前
rms_norm_dynamic_quant	rms_norm_dynamic_quant aclnn文档约束补充&torch_extension中load位置修改 Co-authored-by: LuckySun<sunwenlong8@huawei.com> # message auto-generated for no-merge-commit merge: !7796 merge aclnn_doc into master 整改rms_norm_dynamic_quant算子的文档 Created-by: LuckySun Commit-by: LuckySun Merged-by: cann-robot Description: ## 描述补充aclnnRmsNormDynamicQuant文档中缺失的约束修改在torch_extension中RmsNormDynamicQuant的OpBuilder 的load引入位置修改rms_norm_dynamic_quant算子PTA接口文档的内容格式以及位置 ## 关联的Issue https://gitcode.com/cann/ops-nn/issues/4226 ## 测试 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 norm/rms_norm_dynamic_quant/docs/aclnnRmsNormDynamicQuant.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7796	3 个月前
rms_norm_grad	还原 infer/dgamma 计算的内联写法将 batch_norm_v3 三个 infer 变体的 Compute 与 rms_norm_grad dgamma 计算改回内联实现, 去掉 InferComputeImpl / ComputePreDgamma 两个 helper 间接,使设备侧计算路径更直观; 不改变计算逻辑、计算顺序与内存分配顺序。 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01RDXGVkU6aL6vJjXEPumYHh	2 个月前
rms_norm_grad_quant	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
rms_norm_quant	RmsNormQuant输出空tensor处理报错 Co-authored-by: 任如海<renruhai1@huawei.com> # message auto-generated for no-merge-commit merge: !7957 merge master into master RmsNormQuant输出空tensor处理报错 Created-by: renruhai Commit-by: 任如海 Merged-by: cann-robot Description: ## 描述 <!--在这里详细描述你的改动，包括改动的原因和所采取的方法。--> ## 关联的Issue <!-- 如果这个PR是为了解决特定的Issue，请在这里提供Issue链接。例如：关联Issue #000--> ## 测试 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [x] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7957	2 个月前
rms_norm_quant_v2	arch35: DataCopyExtParams迁移及文档错误码修正 Co-authored-by: raoliang_sac<raoliang4@huawei.com> # message auto-generated for no-merge-commit merge: !7482 merge cherry-pick-mr-7319-1784000275331-auto into master arch35: DataCopyExtParams迁移及文档错误码修正 Created-by: raoliang_sac Commit-by: raoliang_sac Merged-by: cann-robot Description: https://gitcode.com/cann/ops-nn/issues/4015 ## 改动内容 ### 1. DataCopyParams → DataCopyExtParams 迁移解决 uint16 在大 UB（如 V121 384KB）下溢出问题，blockLen/srcStride/dstStride 升级为 uint32。涉及 arch35 kernel 文件： - `add_rms_norm_dynamic_mx_quant` arch35 (fp8/fp4 split_r) - `rms_norm_quant_v2` arch35 (regbase_recompute) - `rms_norm_dynamic_mx_quant` arch35 (recompute) 配套适配：`DataCopyPadParams` → `DataCopyPadExtParams<T_X>` ### 2. C-style cast 修复 `add_rms_norm_quant_tiling_arch35.cpp` 中 `(int64_t)` → `static_cast<int64_t>`，符合 G.EXP.14-CPP 规范 ### 3. 文档错误码修正 3 个 rmsnorm+quant 算子文档中 `ACLNN_ERR_INNER_TILING_ERROR(561002)` 改为 `ACLNN_ERR_INNER_NULLPTR(561103)`： - `add_rms_norm_dynamic_mx_quant` - `rms_norm_dynamic_mx_quant` - `rms_norm_quant_v3` ### 4. 修复 GCC 14.2 ICE 参考 https://gitcode.com/cann/community/issues/157 ，将 `castTrait` 定义从模板函数内部 `static constexpr` 提取到命名空间作用域，规避 GCC 14.2 模板实例化时 ICE： - `add_rms_norm_dynamic_mx_quant_common.h` - `rms_norm_dynamic_mx_quant_common.h` ## 涉及文件（10个，全部 arch35） \| 文件 \| 改动内容 \| \|------\|--------\| \| add_rms_norm_dynamic_mx_quant arch35 common.h \| GCC14 ICE 修复 + DataCopyExtParams迁移 \| \| add_rms_norm_dynamic_mx_quant arch35 fp8_split_r.h \| DataCopyExtParams迁移 \| \| add_rms_norm_dynamic_mx_quant arch35 fp4_split_r.h \| DataCopyExtParams迁移 \| \| add_rms_norm_dynamic_mx_quant docs \| 错误码修正 \| \| add_rms_norm_quant tiling_arch35.cpp \| C-style cast修复 \| \| rms_norm_dynamic_mx_quant arch35 common.h \| GCC14 ICE 修复 \| \| rms_norm_dynamic_mx_quant arch35 recompute.h \| DataCopyExtParams迁移 \| \| rms_norm_dynamic_mx_quant docs \| 错误码修正 \| \| rms_norm_quant_v2 arch35 regbase_recompute.h \| DataCopyExtParams迁移 \| \| rms_norm_quant_v3 docs \| 错误码修正 \| See merge request: cann/ops-nn!7482	3 个月前
rms_norm_quant_v3	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
sync_batch_norm_backward_elemt	add cases Co-authored-by: zhaogrug<zhaohai16@h-partners.com> # message auto-generated for no-merge-commit merge: !8026 merge the_rest_of_cases into master add cases Created-by: zhaogrug Commit-by: zhaogrug Merged-by: cann-robot Description: ## 描述新增仓内新版ttk用例 <!--在这里详细描述你的改动，包括改动的原因和所采取的方法。--> ## 关联的Issue [#4386](https://gitcode.com/cann/ops-nn/issues/4386) <!-- 如果这个PR是为了解决特定的Issue，请在这里提供Issue链接。例如：关联Issue #000--> ## 测试冒烟测试，本地验证 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [x] 其他，请描述：用例添加 ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!8026	2 个月前
sync_batch_norm_backward_reduce	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
sync_batch_norm_gather_stats	fix: comment out dlog_setlevel calls in UT tests for target operators Co-authored-by: liangyuhua<liangyuhua4@huawei.com> # message auto-generated for no-merge-commit merge: !8042 merge fix/comment-out-dlog-setlevel into master fix: comment out dlog_setlevel calls in UT tests for target operators Created-by: liangyuhua Commit-by: liangyuhua Merged-by: cann-robot Description: ## 描述删除 11 个目标算子 UT 测试文件中所有 `dlog_setlevel` 调用（含已注释的），避免 UT 运行时修改全局日志级别影响其他用例，同时清理无用注释代码。修复 `batch_norm_grad_v3` 中变量名 `workspace_size_holer` 拼写错误为 `workspace_size_holder`。涉及的算子及文件： \| 模块 \| 算子 \| 文件 \| 删除数量 \| \|------\|------\|------\|---------\| \| activation \| ConfusionSoftmaxGrad \| test_confusion_softmax_grad_tiling.cpp \| 6 \| \| activation \| LogSoftmaxGrad \| test_log_softmax_grad_tiling.cpp \| 14 \| \| activation \| SoftmaxGrad \| test_softmax_grad_tiling.cpp \| 12 \| \| index \| BroadcastGradientArgs \| test_broadcast_gradient_args_tiling.cpp \| 32 \| \| norm \| BatchNorm \| test_batch_norm_tiling.cpp \| 3 \| \| norm \| BatchNormGrad \| test_batch_norm_grad_tiling.cpp \| 48 \| \| norm \| BatchNormGradV3 \| test_batch_norm_grad_v3_tiling.cpp \| 44 \| \| norm \| LayerNorm \| test_layer_norm_tiling.cpp \| 2 \| \| norm \| LayerNormV3 \| test_layer_norm_v3_tiling.cpp \| 8 \| \| norm \| RmsNormDynamicMxQuant \| test_rms_norm_dynamic_mx_quant_infershape.cpp \| 2 \| \| norm \| SyncBatchNormGatherStats \| test_sync_batch_norm_gather_stats_tiling.cpp \| 12 \| 共计 11 个文件，删除 183 处 `dlog_setlevel` 调用。额外修复： - `batch_norm_grad_v3` 文件中 46 处变量名 `workspace_size_holer` 拼写错误修正为 `workspace_size_holder` - `layer_norm_v3` 文件 license header 从旧格式更新为标准格式注：RmsNormGrad、SoftmaxV2、LogSoftmaxV2 的 `dlog_setlevel` 删除已通过前置 PR 合入 master，本 PR 不再重复修改。 ## 关联的Issue #4396 ## 测试 - [x] pre-commit 全部通过（clang-format + OAT + codespell 等，11 个文件手动扫描确认） - [x] 仅删除 `dlog_setlevel` 调用及修复变量名拼写，不影响测试逻辑 ## 文档更新无需更新文档 ## 类型标签 - [x] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 - [x] AI辅助编写 See merge request: cann/ops-nn!8042	2 个月前
sync_batch_norm_gather_stats_fused	modified md file(aclnn.md and README.md) Co-authored-by: duhuiping<duhuiping@h-partners.com> # message auto-generated for no-merge-commit merge: !7119 merge master into master modified md file(aclnn.md and README.md) Created-by: gitee-duhuiping Commit-by: duhuiping Merged-by: cann-robot Description: ## 描述 1. Updated the constraints on the C and group parameters in aclnnGroupNormBackward.md. 2. Updated the calculation formula and parameter types of the lp_norm_v2 operator, and updated the parameter description of aclnn. 3. Updated the case of parameter names in the README.md file of renorm. 4. Updated the calculation formula of sync_batch_norm_backward_elemt and added the constraints on the combination of data types in the README file. 5. Updated the attribute parameter types and default values in the README file of sync_batch_norm_backward_elemt. ## 关联的Issue [#3701](https://gitcode.com/cann/ops-nn/issues/3701) [#3702](https://gitcode.com/cann/ops-nn/issues/3702) [#3693](https://gitcode.com/cann/ops-nn/issues/3693) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 docs/zh/op_list.md norm/group_norm_grad/docs/aclnnGroupNormBackward.md norm/lp_norm_v2/README.md norm/lp_norm_v2/docs/aclnnNorm.md norm/renorm/README.md norm/sync_batch_norm_backward_elemt/README.md norm/sync_batch_norm_backward_elemt/docs/aclnnBatchNormElemtBackward.md norm/sync_batch_norm_backward_elemt/docs/aclnnBatchNormElemtBackward.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7119	3 个月前
sync_batch_norm_gather_stats_with_counts	modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Co-authored-by: zhaozhoujun520<zhaozhoujun@huawei.com> # message auto-generated for no-merge-commit merge: !7858 merge master into master modified md files (The product filtering tag is added to the aclnn.md files of the forech and norm classes) Created-by: gitee-duhuiping Commit-by: zhaozhoujun520 Merged-by: cann-robot Description: ## 描述 The product filtering tag is added to the aclnn.md files of the forech and norm classes. ## 关联的Issue [#4263](https://gitcode.com/cann/ops-nn/issues/4263) ## 测试 Only involves updating the MD document description, does not involve testing. ## 文档更新 aclnn.md ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [x] 文档更新 - [ ] 其他，请描述： ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!7858	3 个月前
sync_bn_training_update	format cpp Co-authored-by: yang-di52<yangdi52@huawei.com> # message auto-generated for no-merge-commit merge: !6784 merge issue_fix into master format cpp Created-by: yang-di52 Commit-by: yang-di52 Merged-by: cann-robot Description: ## 描述批量刷新cpp代码格式 ## 关联的Issue [#3791](https://gitcode.com/cann/ops-nn/issues/3791) ## 测试 <!--描述进行了哪些测试来验证你的改动。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [x] 其他，请描述：代码格式化 ## AI/Agent生成声明 <!-- [x] 表示选中 --> - [ ] AI辅助编写 See merge request: cann/ops-nn!6784	3 个月前
CMakeLists.txt	Format update: Adjusted the license description format in all relevant files. Co-authored-by: jiangqi<jiangqi47@h-partners.com> # message auto-generated for no-merge-commit merge: !365 merge license5 into master Format update: Adjusted the license description format in all relevant files. Created-by: jiangqi2025 Commit-by: jiangqi Merged-by: turing_project1 Description: ## 描述 <!--在这里详细描述你的改动，包括改动的原因和所采取的方法。--> ## 关联的Issue <!-- 如果这个PR是为了解决特定的Issue，请在这里提供Issue链接。例如：关联Issue #123--> <!-- 如果这个PR是为了解决特定的问题单，请在这里描述问题单单号。--> ## 测试 <!--描述进行了哪些测试来验证你的改动。包括但不限于二级冒烟、算子泛化等。--> ## 文档更新 <!--如果这个PR包含文档的更新，请在这里指出。例如：更新了README.md文件。--> ## 类型标签 <!-- [x] 表示选中 --> - [ ] Bug修复 - [ ] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： See merge request: cann/ops-nn!365	10 个月前