op-plugin/op_plugin/ops/opapi · wangqi_ai/op-plugin

文件	最后提交记录	最后更新时间
sparse	!2233 Fix cleancode. Merge pull request !2233 from 王夏夏/master	1 年前
AdaptiveAvgPool2dKernelNpuOpApi.cpp	!2210 [Fix] Fix special case and add check for adaptive_avg_pool2d. Merge pull request !2210 from 刘嘉巍/dev	1 年前
AdaptiveMaxPool3dBackwardKernelNpuOpApi.cpp	!2488 Refactor compability judgement. Merge pull request !2488 from 刘嘉巍/master	1 年前
AdaptiveMaxPool3dKernelNpuOpApi.cpp	!2488 Refactor compability judgement. Merge pull request !2488 from 刘嘉巍/master	1 年前
AddKernelNpuOpApi.cpp	!2748 Fuzz fix add and addr Merge pull request !2748 from wgb/bugfix	1 年前
AddLayerNormBackwardKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
AddLayerNormKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
AddRmsNormQuantKernelOpApi.cpp	fix addrmsnormquant output Co-authored-by: Zichun Ye<zichun.ye@huawei.com> # message auto-generated for no-merge-commit merge: merge rmsnorm_out_fix into master fix addrmsnormquant output Created-by: zichun_ye Commit-by: Zichun Ye Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: To adapter the change in aclnn operator, we add a dummy output rmsnorm_out in the adapter codes. PTA doesn't support this output for now. Special notes for your reviewers: See merge request: Ascend/op-plugin!3282	1 年前
AddbmmKernelNpuOpApi.cpp	!2265 fix cleancode Merge pull request !2265 from 詹昊/cleancode1	1 年前
AddcdivKernelNpuOpApi.cpp	!2770 fix fuzz addcdiv Merge pull request !2770 from zhangqiongwen/master_addcdiv_fuzz	1 年前
AddcmulKernelNpuOpApi.cpp	!2542 fix bug when addcmul broadcast Merge pull request !2542 from zhangqiongwen/master_addcmul_bug_fix	1 年前
AddmmKernelNpuOpApi.cpp	!2865 开启aclnnAddmm weightNz功能，op-plugin侧修改 Merge pull request !2865 from liuchengzhuo/enable_addmm_weightNz	1 年前
AddmvKernelNpuOpApi.cpp	!2204 clean code : infershape file Merge pull request !2204 from wgb/clean_code	1 年前
AdvanceStepKernelNpuOpApi.cpp	!3031 add AdvanceStep to support speculative inference Merge pull request !3031 from 陈赵旻熠/AdvanceStep	1 年前
AffineGridGeneratorBackwardKernelNpuOpApi.cpp	!3084 [Feature] aclnn for affine_grid_generator_backward Merge pull request !3084 from 刘嘉巍/master	1 年前
AffineGridGeneratorKernelNpuOpApi.cpp	!2302 [5/N] cleancode Merge pull request !2302 from zhangqiongwen/master_cleancode_05	1 年前
AllGatherBaseMatmulKernelNpuOpApi.cpp	npu_all_gather_base_mm support quant and comm_mode aiv Co-authored-by: wang-guangbin<wgb_strive@163.com> # message auto-generated for no-merge-commit merge: merge allgather into master npu_all_gather_base_mm support quant and comm_mode aiv Created-by: wang-guangbin Commit-by: wang-guangbin Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: 后续添加测试用例和API资料。当前用例本地功能ok，精度验证时框架IPC功能存在问题，后续解决上库UT。 See merge request: Ascend/op-plugin!3224	1 年前
AllKernelNpuOpApi.cpp	!2260 fix cleancode Merge pull request !2260 from 詹昊/master	1 年前
AllToAllvGmmNpuOpApi.cpp	gmmalltoallv & alltoallvgmm fix hcom_ptr Co-authored-by: MeiWenxuan<meiwenxuan@huawei.com> # message auto-generated for no-merge-commit merge: !4141 merge 7.3.0 into 7.3.0 gmmalltoallv & alltoallvgmm fix hcom_ptr Created-by: MeiWenxuan Commit-by: MeiWenxuan Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!4141	8 个月前
AminmaxKernelNpuOpApi.cpp	!2233 Fix cleancode. Merge pull request !2233 from 王夏夏/master	1 年前
AmpForeachNonFiniteCheckAndUnscaleKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
AmpUpdateScaleKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (5/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge dev into master Delete code for v1.11.0 and v2.0 (5/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3187	1 年前
AntiQuantKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
AnyKernelNpuOpApi.cpp	!2302 [5/N] cleancode Merge pull request !2302 from zhangqiongwen/master_cleancode_05	1 年前
ApplyRotaryPosEmbNpuOpApi.cpp	support tnd format Co-authored-by: ghost<zhengwenbang@huawei.com> # message auto-generated for no-merge-commit merge: !3355 merge master into master support tnd format Created-by: weixin_44564637 Commit-by: ghost Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3355	1 年前
ArangeKernelNpuOpApi.cpp	!2295 [3/N]cleancode Merge pull request !2295 from zhangqiongwen/master_cleancode_0318_01	1 年前
ArgSortKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
ArgmaxKernelNpuOpApi.cpp	fix: when dim=none keepdim=true, PTA and native interface outputs are inconsistent. Co-authored-by: culechan<chenshuai81@huawei.com> # message auto-generated for no-merge-commit merge: !3520 merge cherry-pick-mr-3453-1763534028619-auto into 7.3.0 fix: when dim=none keepdim=true, PTA and native interface outputs are inconsistent. Created-by: culechan Commit-by: culechan Merged-by: ascend-robot Description: torch.argmax接口在dim未指定且keepdim=true的时候，npu与原生框架表现不一致。原生框架输出为张量，npu为标量 See merge request: Ascend/op-plugin!3520	10 个月前
ArgminKernelNpuOpApi.cpp	!2275 fix cleancode Merge pull request !2275 from 詹昊/cleancode1	1 年前
Atan2KernelNpuOpApi.cpp	!1666 [Feature] Codegen for multi-named tensor. Merge pull request !1666 from 刘嘉巍/master-add	1 年前
AttentionUpdateKernelNpuOpApi.cpp	Add npu_attention_update interface Co-authored-by: huawuyi<liuxuehui4@huawei.com> # message auto-generated for no-merge-commit merge: merge master into master Add npu_attention_update interface Created-by: huawuyi Commit-by: huawuyi Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3166	1 年前
AttentionWorkerCombineOpApi.cpp	Add AttentionWorkerCombine PTA Co-authored-by: weixin_36491311<lingwanxin@huawei.com> # message auto-generated for no-merge-commit merge: merge master into master Add AttentionWorkerCombine PTA Created-by: weixin_36491311 Commit-by: weixin_36491311 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3176	1 年前
AttentionWorkerSchedulerKernelNpuOpApi.cpp	适配支持两类算子：ffn_worker_scheduler_/ffn_worker_scheduler、 attention_worker_scheduler_/attention_worker_scheduler Co-authored-by: panweicheng<panweicheng@huawei.com> # message auto-generated for no-merge-commit merge: !3397 merge master into master 适配支持两类算子：ffn_worker_scheduler_/ffn_worker_scheduler、 attention_worker_scheduler_/attention_worker_scheduler Created-by: pwc10490 Commit-by: panweicheng Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > /kind feature What does this PR do / why do we need it: 适配支持两类算子： torch_npu.attention_worker_scheduler torch_npu.attention_worker_scheduler_（inplace算子） torch_npu.ffn_worker_scheduler torch_npu.ffn_worker_scheduler_（inplace算子） Special notes for your reviewers: See merge request: Ascend/op-plugin!3397	11 个月前
AttnSoftMaxBackwardKernelNpuOpApi.cpp	support undocumented cube_math_type Co-authored-by: freezee<gengdongjie@huawei.com> # message auto-generated for no-merge-commit merge: !3834 merge 7.3.0 into 7.3.0 support undocumented cube_math_type Created-by: freezee Commit-by: freezee Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3834	9 个月前
AvgPool2dBackwardKernelNpuOpApi.cpp	!2204 clean code : infershape file Merge pull request !2204 from wgb/clean_code	1 年前
AvgPool2dKernelNpuOpApi.cpp	!2302 [5/N] cleancode Merge pull request !2302 from zhangqiongwen/master_cleancode_05	1 年前
AvgPool3dBackwardKernelNpuOpApi.cpp	!1845 增加avgpool3dbackward的路径五调用 Merge pull request !1845 from 唐燕锋/avgpool3d	2 年前
AvgPool3dKernelNpuOpApi.cpp	!2311 modify cleancode/9 Merge pull request !2311 from 史佳馨/master	1 年前
BaddbmmKernelNpuOpApi.cpp	!2276 fix cleancode Merge pull request !2276 from 詹昊/cleancode2	1 年前
BatchGatherMatmulKernelNpuOpApi.cpp	!2274 add npu_batch_gather_matmul inplace Merge pull request !2274 from 沈珈靓/master	1 年前
BatchNormBackwardKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
BatchNormBackwardReduceKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
BatchNormElemtKernelNpuOpApi.cpp	!2265 fix cleancode Merge pull request !2265 from 詹昊/cleancode1	1 年前
BatchNormKernelOpApi.cpp	!3037 Removed the overfitting of _batch_norm_impl_index, _batch_norm_impl_index_backward, and batch_norm. Merge pull request !3037 from 史佳馨/master	1 年前
BernoulliKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
BincountKernelNpuOpApi.cpp	!2328 cleancode Merge pull request !2328 from DaiFu/2503CleanCode2	1 年前
BitwiseAndKernelNpuOpApi.cpp	!2245 [Codecheck] Fix codecheck. Merge pull request !2245 from 刘嘉巍/x1	1 年前
BitwiseOrKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
BitwiseXorKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
BmmKernelNpuOpApi.cpp	Check the input dims of bmm function. Co-authored-by: zhujiaxing<zhujiaxing2@huawei.com> # message auto-generated for no-merge-commit merge: !3437 merge master into master Check the input dims of bmm function. Created-by: zhujiaxing1029 Commit-by: zhujiaxing Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? /kind bug What does this PR do / why do we need it: Check the input dims of bmm function. Special notes for your reviewers: See merge request: Ascend/op-plugin!3437	11 个月前
BucketizeKernelNpuOpApi.cpp	!2247 cleancode for ops. Merge pull request !2247 from 王夏夏/master	1 年前
CastKernelNpuOpApi.cpp	!2601 Add TensorWrapper and _npu_dtype_cast Merge pull request !2601 from chuboning/master	1 年前
CatKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (5/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge dev into master Delete code for v1.11.0 and v2.0 (5/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3187	1 年前
ComplexKernelNpuOpApi.cpp	!2326 [Task] cleancode 2/N. Merge pull request !2326 from will-devil/cleancode-1	1 年前
ConstantPadNdKernelNpuOpApi.cpp	!2276 fix cleancode Merge pull request !2276 from 詹昊/cleancode2	1 年前
ConvDepthwise2dKernelNpuOpApi.cpp	!2149 TORCH MAIN SYNC : remove use_const_ref_for_mutable_tensors Merge pull request !2149 from dilililiwhy/main_sync_20250125	1 年前
ConvertWeightToINT4PackKernelNpuOpApi.cpp	!1994 int4pack接口支持nz输入 Merge pull request !1994 from zihan007/master	1 年前
ConvolutionBackwardKernelNpuOpApi.cpp	[revert] fix conv3d mask Co-authored-by: shaoyf<shaoyifan1@huawei.com> # message auto-generated for no-merge-commit merge: !4156 merge revert-mr-3998-1769825255340-auto into 7.3.0 [revert] fix conv3d mask Created-by: shaoyf Commit-by: shaoyf Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!4156	6 个月前
ConvolutionKernelNpuOpApi.cpp	mod conv1d2d Co-authored-by: shaoyf<shaoyifan1@huawei.com> # message auto-generated for no-merge-commit merge: !3775 merge revert-mr-3705-1765460870798-auto into 7.3.0 mod conv1d2d Created-by: shaoyf Commit-by: shaoyf Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3775	10 个月前
CountNonZeroKernelNpuOpApi.cpp	!2217 fix cleancode Merge pull request !2217 from 詹昊/cleancode8	1 年前
CtcLossKernelNpuOpApi.cpp	!2988 remove dedundant ctc_loss Merge pull request !2988 from DaiFu/2508ctc	1 年前
CummaxKernelNpuOpApi.cpp	!2326 [Task] cleancode 2/N. Merge pull request !2326 from will-devil/cleancode-1	1 年前
CumminKernelNpuOpApi.cpp	!2326 [Task] cleancode 2/N. Merge pull request !2326 from will-devil/cleancode-1	1 年前
CumsumKernelNpuOpApi.cpp	!2282 Fix codecheck. Merge pull request !2282 from 刘嘉巍/cleancode	1 年前
DeepNormBackwardKernelOpApi.cpp	!1713 fix ops * fix ops	2 年前
DeepNormKernelOpApi.cpp	!1713 fix ops * fix ops	2 年前
DequantBiasNpuOpApi.cpp	!2188 dequant_bias Merge pull request !2188 from liushiyu0214/master	1 年前
DequantRopeQuantKVCacheKernelNpuOpApi.cpp	!2103 modify op dequant_repo_quant_kv_cache file name Merge pull request !2103 from 流浪者LingLong/master	1 年前
DequantSwigluQuantOpApi.cpp	!3111 add swiglu_mode,clamp_limit,glu_alpha,glu_bias attr for DequantSwigluQuant Merge pull request !3111 from ghost/master	1 年前
DistributeBarrierOpApi.cpp	barrier Co-authored-by: davon<liuwenda4@huawei.com> # message auto-generated for no-merge-commit merge: !3372 merge master into master barrier Created-by: Davon14272 Commit-by: davon Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind feature What does this PR do / why do we need it: _npu_distribute_barrier adds two optional inputs: time_out and elastic_info Special notes for your reviewers: See merge request: Ascend/op-plugin!3372	11 个月前
DivKernelNpuOpApi.cpp	fix cast logic in div Co-authored-by: Zichun Ye<zichun.ye@huawei.com> # message auto-generated for no-merge-commit merge: merge div_cast_fix into master fix cast logic in div Created-by: zichun_ye Commit-by: Zichun Ye Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Change the original logic in div to fit that of div_out and original torch. Special notes for your reviewers: See merge request: Ascend/op-plugin!3300	1 年前
DropoutGenMaskKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
DropoutKernelNpuOpApi.cpp	!1757 [Feature] Support dropout inplace. Merge pull request !1757 from will-devil/dropout_inplace	2 年前
DynamicBlockQuantNpuOpApi.cpp	!3059 [Feat] Add npu_dynamic_block_quant adpt Merge pull request !3059 from zhangqiongwen/master_DynamicBlockQuant	1 年前
DynamicQuantKernelNpuOpApi.cpp	Remove unused GE apis Co-authored-by: li_jing_hw<lijing202@huawei.com> # message auto-generated for no-merge-commit merge: merge master into master Remove unused GE apis Created-by: li_jing_hw Commit-by: li_jing_hw Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 相关功能属于废弃功能，当前不使用，同时删除无关的GE依赖 Special notes for your reviewers: See merge request: Ascend/op-plugin!3210	1 年前
EmbeddingBagNpuOpApi.cpp	!3079 [Feature] aclnn for several apis Merge pull request !3079 from 刘嘉巍/meta	1 年前
EmbeddingKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
EmbeddingRenormKernelNpuOpApi.cpp	!2311 modify cleancode/9 Merge pull request !2311 from 史佳馨/master	1 年前
EqKernelNpuOpApi.cpp	!1666 [Feature] Codegen for multi-named tensor. Merge pull request !1666 from 刘嘉巍/master-add	1 年前
EqualKernelNpuOpApi.cpp	!1666 [Feature] Codegen for multi-named tensor. Merge pull request !1666 from 刘嘉巍/master-add	1 年前
ExponentialKernelNpuOpApi.cpp	!2247 cleancode for ops. Merge pull request !2247 from 王夏夏/master	1 年前
EyeKernelNpuOpApi.cpp	!2245 [Codecheck] Fix codecheck. Merge pull request !2245 from 刘嘉巍/x1	1 年前
FFNKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
FFTShiftKernelNpuOpApi.cpp	!2353 [Task] cleancode 3/N. Merge pull request !2353 from will-devil/cleancode-2	1 年前
FFTc2cKernelNpuOpApi.cpp	!2791 fix _fft_r2c_out Merge pull request !2791 from shaoyf/test_fft	1 年前
FFTc2rKernelNpuOpApi.cpp	!1836 custimize backward for c2r and r2c to reduce aicpu calls Merge pull request !1836 from alanhe151220037/master	2 年前
FFTr2cKernelNpuOpApi.cpp	!1836 custimize backward for c2r and r2c to reduce aicpu calls Merge pull request !1836 from alanhe151220037/master	2 年前
FakeQuantizeKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
FastGeluNpuOpApi.cpp	[fix] Add Ascend chip type judgment for fast_gelu Co-authored-by: liangsongwei<liangsongwei@huawei.com> # message auto-generated for no-merge-commit merge: !3699 merge 7.3.0 into 7.3.0 [fix] Add Ascend chip type judgment for fast_gelu Created-by: liangsongwei Commit-by: liangsongwei Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3699	10 个月前
FfnWorkerBatchingKernelNpuOpApi.cpp	add FfnWorkerBatching Co-authored-by: 安冬<andong2@huawei.com> # message auto-generated for no-merge-commit merge: merge master into master add FfnWorkerBatching Created-by: andong_hw Commit-by: 安冬 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3179	1 年前
FfnWorkerSchedulerKernelNpuOpApi.cpp	适配支持两类算子：ffn_worker_scheduler_/ffn_worker_scheduler、 attention_worker_scheduler_/attention_worker_scheduler Co-authored-by: panweicheng<panweicheng@huawei.com> # message auto-generated for no-merge-commit merge: !3397 merge master into master 适配支持两类算子：ffn_worker_scheduler_/ffn_worker_scheduler、 attention_worker_scheduler_/attention_worker_scheduler Created-by: pwc10490 Commit-by: panweicheng Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > /kind feature What does this PR do / why do we need it: 适配支持两类算子： torch_npu.attention_worker_scheduler torch_npu.attention_worker_scheduler_（inplace算子） torch_npu.ffn_worker_scheduler torch_npu.ffn_worker_scheduler_（inplace算子） Special notes for your reviewers: See merge request: Ascend/op-plugin!3397	11 个月前
FftRfftKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
FillKernelNpuOpApi.cpp	!2311 modify cleancode/9 Merge pull request !2311 from 史佳馨/master	1 年前
FlashAttentionKernelNpuOpApi.cpp	[feat]]add sink for fa Co-authored-by: zqwenn<zhangqiongwen@huawei.com> # message auto-generated for no-merge-commit merge: !3888 merge 7.3.0_fav5_apiv1 into 7.3.0 [feat]]add sink for fa Created-by: zqwenn Commit-by: zqwenn Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3888	9 个月前
FlashAttentionV2KernelNpuOpApi.cpp	[feat]npu_fusion_attention_v2 adapt for fav5 Co-authored-by: zqwenn<zhangqiongwen@huawei.com> # message auto-generated for no-merge-commit merge: !3522 merge 7.3.0_fav5_adpt into 7.3.0 [feat]npu_fusion_attention_v2 adapt for fav5 Created-by: zqwenn Commit-by: zqwenn Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: npu_fusion_attention_v2新增softmax_layout和sink入参新增场景测试用例下一个pr提交 See merge request: Ascend/op-plugin!3522	10 个月前
FloorDivideKernelNpuOpApi.cpp	!1666 [Feature] Codegen for multi-named tensor. Merge pull request !1666 from 刘嘉巍/master-add	1 年前
ForeachAbsKernelOpApi.cpp	!2317 Cleancode for ops. Merge pull request !2317 from 王夏夏/cleancode	1 年前
ForeachAcosKernelNpuOpApi.cpp	!2309 fix cleancode Merge pull request !2309 from 詹昊/master	1 年前
ForeachAddKernelNpuOpApi.cpp	!2245 [Codecheck] Fix codecheck. Merge pull request !2245 from 刘嘉巍/x1	1 年前
ForeachAddScalarKernelNpuOpApi.cpp	!2246 [Codecheck] Fix codecheck. Merge pull request !2246 from 刘嘉巍/deformable	1 年前
ForeachAddcdivScalarKernelNpuOpApi.cpp	!2316 cleancode Merge pull request !2316 from DaiFu/2503CleanCode1	1 年前
ForeachAddcdivScalarListKernelNpuOpApi.cpp	!2282 Fix codecheck. Merge pull request !2282 from 刘嘉巍/cleancode	1 年前
ForeachAddcdivTensorKernelNpuOpApi.cpp	!2411 fix foreach addcdiv(/addcmul).Tensor tesnor slice bug Merge pull request !2411 from 流浪者LingLong/master	1 年前
ForeachAddcmulScalarKernelNpuOpApi.cpp	!2205 cleancode for ops. Merge pull request !2205 from 王夏夏/master	1 年前
ForeachAddcmulScalarListKernelNpuOpApi.cpp	!1868 foreach_addcdiv/mul.scalarlist, foreach_addcdiv/mul.Tensor代码回退 Merge pull request !1868 from zhanghao/master	2 年前
ForeachAddcmulTensorKernelNpuOpApi.cpp	!2411 fix foreach addcdiv(/addcmul).Tensor tesnor slice bug Merge pull request !2411 from 流浪者LingLong/master	1 年前
ForeachAsinKernelNpuOpApi.cpp	!2215 [Codecheck] Fix codecheck. Merge pull request !2215 from 刘嘉巍/x1	1 年前
ForeachAtanKernelNpuOpApi.cpp	!2311 modify cleancode/9 Merge pull request !2311 from 史佳馨/master	1 年前
ForeachClampMaxKernelNpuOpApi.cpp	!2276 fix cleancode Merge pull request !2276 from 詹昊/cleancode2	1 年前
ForeachCopyKernelOpApi.cpp	batch copy en queue Co-authored-by: glf2016<13739288457@163.com> # message auto-generated for no-merge-commit merge: !3229 merge queue into master batch copy en queue Created-by: glf2016 Commit-by: glf2016 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3229	1 年前
ForeachCosKernelNpuOpApi.cpp	!2328 cleancode Merge pull request !2328 from DaiFu/2503CleanCode2	1 年前
ForeachCoshKernelNpuOpApi.cpp	!2309 fix cleancode Merge pull request !2309 from 詹昊/master	1 年前
ForeachDivKernelNpuOpApi.cpp	!2218 [1/N]cleancode Format Merge pull request !2218 from zhangqiongwen/master_cleancode_01	1 年前
ForeachErfKernelNpuOpApi.cpp	!2311 modify cleancode/9 Merge pull request !2311 from 史佳馨/master	1 年前
ForeachErfcKernelNpuOpApi.cpp	!2005 foreach op max/min list/scalarlist Data type compatibility Merge pull request !2005 from 流浪者LingLong/master	1 年前
ForeachExpKernelNpuOpApi.cpp	!2328 cleancode Merge pull request !2328 from DaiFu/2503CleanCode2	1 年前
ForeachExpm1KernelNpuOpApi.cpp	!2233 Fix cleancode. Merge pull request !2233 from 王夏夏/master	1 年前
ForeachLerpListKernelNpuOpApi.cpp	!2317 Cleancode for ops. Merge pull request !2317 from 王夏夏/cleancode	1 年前
ForeachLerpScalarKernelNpuOpApi.cpp	!2317 Cleancode for ops. Merge pull request !2317 from 王夏夏/cleancode	1 年前
ForeachLog10KernelNpuOpApi.cpp	!2311 modify cleancode/9 Merge pull request !2311 from 史佳馨/master	1 年前
ForeachLog1pKernelNpuOpApi.cpp	!2309 fix cleancode Merge pull request !2309 from 詹昊/master	1 年前
ForeachLog2KernelNpuOpApi.cpp	!2309 fix cleancode Merge pull request !2309 from 詹昊/master	1 年前
ForeachLogKernelNpuOpApi.cpp	!1800 Added foreach op erf erfc exp expm1 floor frac log log1p log10 mullist mulscalarlist compatibility processing &&Modified ut Merge pull request !1800 from 流浪者LingLong/master	2 年前
ForeachMaximumKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
ForeachMinimumKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
ForeachMulKernelNpuOpApi.cpp	!2235 cleancode Merge pull request !2235 from DaiFu/2503CleanCode	1 年前
ForeachMulScalarKernelOpApi.cpp	!1822 新增foreach部分算子V2接口 Merge pull request !1822 from 王育才/master	2 年前
ForeachNegKernelNpuOpApi.cpp	!2215 [Codecheck] Fix codecheck. Merge pull request !2215 from 刘嘉巍/x1	1 年前
ForeachNormKernelOpApi.cpp	fix foreach norm out shape bug Co-authored-by: tramp-ll<linglong2@huawei.com> # message auto-generated for no-merge-commit merge: merge master into master fix foreach norm out shape bug Created-by: tramp-ll Commit-by: tramp-ll Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: fix foreach norm out shape bug 修补foreach_norm 的out shape的你npu 与竞品不一致情形：竞品out tensor的shape为0维，而npu的out tensor为1维。与cann变更无关 What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3181	1 年前
ForeachPowKernelNpuOpApi.cpp	!2260 fix cleancode Merge pull request !2260 from 詹昊/master	1 年前
ForeachPowScalarAndTensorKernelNpuOpApi.cpp	!2233 Fix cleancode. Merge pull request !2233 from 王夏夏/master	1 年前
ForeachPowScalarKernelNpuOpApi.cpp	!2323 modify cleancode_3/2 Merge pull request !2323 from 史佳馨/master	1 年前
ForeachReciprocalKernelNpuOpApi.cpp	!2215 [Codecheck] Fix codecheck. Merge pull request !2215 from 刘嘉巍/x1	1 年前
ForeachRoundOffNumberKernelNpuOpApi.cpp	!2316 cleancode Merge pull request !2316 from DaiFu/2503CleanCode1	1 年前
ForeachSigmoidKernelOpApi.cpp	!1802 add foreach op reciprocal powlist powscalarlist powsclarandtensor round sin sigmoid sin sinh sqrt sub_sclarlist tan tanh zero pow_sclar sub_list sub_scalar Compatibility processing Merge pull request !1802 from 流浪者LingLong/master	2 年前
ForeachSignKernelNpuOpApi.cpp	!2401 foreach_sign and foreach_copy add support type Merge pull request !2401 from liujie_1234/master	1 年前
ForeachSinKernelNpuOpApi.cpp	!2309 fix cleancode Merge pull request !2309 from 詹昊/master	1 年前
ForeachSinhKernelNpuOpApi.cpp	!2316 cleancode Merge pull request !2316 from DaiFu/2503CleanCode1	1 年前
ForeachSqrtKernelNpuOpApi.cpp	!1802 add foreach op reciprocal powlist powscalarlist powsclarandtensor round sin sigmoid sin sinh sqrt sub_sclarlist tan tanh zero pow_sclar sub_list sub_scalar Compatibility processing Merge pull request !1802 from 流浪者LingLong/master	2 年前
ForeachSubKernelNpuOpApi.cpp	!1822 新增foreach部分算子V2接口 Merge pull request !1822 from 王育才/master	2 年前
ForeachSubScalarInplaceKernelOpApi.cpp	!1822 新增foreach部分算子V2接口 Merge pull request !1822 from 王育才/master	2 年前
ForeachTanKernelNpuOpApi.cpp	!2215 [Codecheck] Fix codecheck. Merge pull request !2215 from 刘嘉巍/x1	1 年前
ForeachTanhKernelNpuOpApi.cpp	!1802 add foreach op reciprocal powlist powscalarlist powsclarandtensor round sin sigmoid sin sinh sqrt sub_sclarlist tan tanh zero pow_sclar sub_list sub_scalar Compatibility processing Merge pull request !1802 from 流浪者LingLong/master	2 年前
ForeachZeroInplaceKernelNpuOpApi.cpp	!1802 add foreach op reciprocal powlist powscalarlist powsclarandtensor round sin sigmoid sin sinh sqrt sub_sclarlist tan tanh zero pow_sclar sub_list sub_scalar Compatibility processing Merge pull request !1802 from 流浪者LingLong/master	2 年前
FusedAdamWKernelNpuOpApi.cpp	!1926 support _fused_adamw_ in all version Merge pull request !1926 from 詹昊/master	1 年前
FusedInferAttentionScoreKernelNpuOpApi.cpp	add _npu_fused_infer_attention_score_infer_output Co-authored-by: yanpengquan<yanpengquan@huawei.com> # message auto-generated for no-merge-commit merge: !3930 merge 7.3.0_fused into 7.3.0 add _npu_fused_infer_attention_score_infer_output Created-by: wangchao430 Commit-by: yanpengquan Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: ![image.png](https://raw.gitcode.com/user-images/assets/7403085/52a37a39-91ad-4342-924f-b9e3f8f6e9cb/image.png 'image.png') See merge request: Ascend/op-plugin!3930	9 个月前
FusedInferAttentionScoreV2KernelNpuOpApi.cpp	support npu_fused_infer_attention_score_v2 tiling update in NPUGraph Co-authored-by: fengye<vinsmoke.feng@huawei.com> # message auto-generated for no-merge-commit merge: !3925 merge 7.3.0 into 7.3.0 support npu_fused_infer_attention_score_v2 tiling update in NPUGraph Created-by: fengye0903 Commit-by: fengye Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: #@unittest.skip( # "Skipping due to outdated CANN version; please update CANN to the latest version and remove this skip") #@SupportedDevices(['Ascend910B']) def test_npu_fused_infer_attention_score_v2_ifa_return_lse_return_softmax_lse_true_with_workspace(self, device="npu"): query = torch.randn(1, 32, 1, 128, dtype=torch.float16).npu() key = torch.randn(1, 32, 2048, 128, dtype=torch.float16).npu() value = torch.randn(1, 32, 2048, 128, dtype=torch.float16).npu() head_dim = 128 return_softmax_lse = True supported_output, lse_out = self.supported_op_exec(query, key, value, head_dim, 1, 32, 1, return_softmax_lse) custom_output = self.custom_op_exec_with_workspace(query, key, value, head_dim, True) attention_output = custom_output[0] softmaxlse_output = custom_output[1] if __name__ == "__main__": run_tests() [root@devserver-hps-feb34b1c-00033 f00941739]# python ./op-plugin/test/test_custom_ops/test_npu_fused_infer_attention_score_v2_with_workspace.py ss. ---------------------------------------------------------------------- Ran 3 tests in 2.006s OK (skipped=2) Special notes for your reviewers: See merge request: Ascend/op-plugin!3925	9 个月前
FusedLinearCrossEntropyLossKernelNpuOpApi.cpp	!3110 Add fused_linear_crocess_ectropy_loss Merge pull request !3110 from shaoyf/fused_linear_0903	1 年前
GatherBackwardKernelNpuOpApi.cpp	!2329 add npu_gather_backward Merge pull request !2329 from 詹昊/gatherbackward	1 年前
GatherKernelNpuOpApi.cpp	!2316 cleancode Merge pull request !2316 from DaiFu/2503CleanCode1	1 年前
GatherSparseBackwardNpuOpApi.cpp	!3085 [Feat] Add npu_gather_sparse_index_backward adpt Merge pull request !3085 from zhangqiongwen/master_npu_gather_sparse_index_backward	1 年前
GeGluKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
GeKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
GeluBackwardKernelNpuOpApi.cpp	torch.nn.functional.gelu preserves api consistency1. Co-authored-by: shi-jiaxin9<shijiaxin10@h-partners.com> # message auto-generated for no-merge-commit merge: !4152 merge 7.3.0 into 7.3.0 torch.nn.functional.gelu preserves api consistency1. Created-by: shi-jiaxin9 Commit-by: shi-jiaxin9 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!4152	8 个月前
GeluKernelNpuOpApi.cpp	torch.nn.functional.gelu preserves api consistency1. Co-authored-by: shi-jiaxin9<shijiaxin10@h-partners.com> # message auto-generated for no-merge-commit merge: !4152 merge 7.3.0 into 7.3.0 torch.nn.functional.gelu preserves api consistency1. Created-by: shi-jiaxin9 Commit-by: shi-jiaxin9 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!4152	8 个月前
GemmaRmsNormKernelOpApi.cpp	!2208 add npu_gemma_rms_norm Merge pull request !2208 from liuyun/master	1 年前
GmmAllToAllvNpuOpApi.cpp	gmmalltoallv & alltoallvgmm fix hcom_ptr Co-authored-by: MeiWenxuan<meiwenxuan@huawei.com> # message auto-generated for no-merge-commit merge: !4141 merge 7.3.0 into 7.3.0 gmmalltoallv & alltoallvgmm fix hcom_ptr Created-by: MeiWenxuan Commit-by: MeiWenxuan Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!4141	8 个月前
GroupNormBackwardKernelNpuOpApi.cpp	!2275 fix cleancode Merge pull request !2275 from 詹昊/cleancode1	1 年前
GroupedMatmulFinalizeRoutingKernelNpuOpApi.cpp	GMMFR adds support for cumsum in groulist and none option in shareInput. Co-authored-by: lilening<lilening@huawei.com> # message auto-generated for no-merge-commit merge: !3581 merge 7.3.0 into 7.3.0 GMMFR adds support for cumsum in groulist and none option in shareInput. Created-by: lilening Commit-by: lilening Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind feature What does this PR do / why do we need it: GMMFR adds support for cumsum in groulist and none option in shareInput. Special notes for your reviewers: See merge request: Ascend/op-plugin!3581	10 个月前
GroupedMatmulKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
GroupedMatmulSwigluQuantNpuOpapi.cpp	!2597 The indexing mode of atb-related enumerated values is modified. Merge pull request !2597 from 史佳馨/master	1 年前
GroupedMatmulSwigluQuantV2NpuOpapi.cpp	[fix]npu_grouped_matmul_swiglu_quant_v2 output shape fix Co-authored-by: zqwenn<zhangqiongwen@huawei.com> # message auto-generated for no-merge-commit merge: !3641 merge 7.3.0_gmm_gelu_quant_v2_fix into 7.3.0 [fix]npu_grouped_matmul_swiglu_quant_v2 output shape fix Created-by: zqwenn Commit-by: zqwenn Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3641	10 个月前
GtKernelNpuOpApi.cpp	!2322 modify cleancode/3_1 Merge pull request !2322 from 史佳馨/master	1 年前
HardSwishBackwardKernelNpuOpApi.cpp	Modify the threshold value of hardswish_backward. Co-authored-by: shi-jiaxin9<shijiaxin10@h-partners.com> # message auto-generated for no-merge-commit merge: merge master into master Modify the threshold value of hardswish_backward. Created-by: shi-jiaxin9 Commit-by: shi-jiaxin9 Merged-by: ascend-robot Description: What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug What does this PR do / why do we need it: Modify the threshold value of hardswish_backward. Special notes for your reviewers: See merge request: Ascend/op-plugin!3203	1 年前
Im2colKernelNpuOpApi.cpp	!2353 [Task] cleancode 3/N. Merge pull request !2353 from will-devil/cleancode-2	1 年前
IndexAddKernelNpuOpApi.cpp	!3086 The aclop for torch._dim_arange, torch._unsafe_index, torch.index_add, and torch.nn.functional.adaptive_avg_pool1d has been switched to aclnn. Merge pull request !3086 from 史佳馨/master	1 年前
IndexFillKernelNpuOpApi.cpp	Modify the index_fill and index_fill_. Co-authored-by: shi-jiaxin9<shijiaxin10@h-partners.com> # message auto-generated for no-merge-commit merge: merge master into master Modify the index_fill and index_fill_. Created-by: shi-jiaxin9 Commit-by: shi-jiaxin9 Merged-by: ascend-robot Description: What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug What does this PR do / why do we need it: Modify the index_fill and index_fill_. Special notes for your reviewers: See merge request: Ascend/op-plugin!3245	1 年前
IndexKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
IndexPutKernelNpuOpApi.cpp	[Revert] index_put_index_put_out_of_bounds_error_interception Co-authored-by: shaoyf<shaoyifan1@huawei.com> # message auto-generated for no-merge-commit merge: !4026 merge revert-mr-4010-1768186894927-auto into 7.3.0 [Revert] index_put_index_put_out_of_bounds_error_interception Created-by: shaoyf Commit-by: shaoyf Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!4026	9 个月前
IndexSelectKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
InterleaveRopeNpuOpApi.cpp	!2781 cleancode Merge pull request !2781 from DaiFu/2507Cleancode	1 年前
IsInKernelNpuOpApi.cpp	!2311 modify cleancode/9 Merge pull request !2311 from 史佳馨/master	1 年前
KthvalueKernelNpuOpApi.cpp	!2246 [Codecheck] Fix codecheck. Merge pull request !2246 from 刘嘉巍/deformable	1 年前
KvQuantSparseFlashAttentionKernelNpuOpApi.cpp	QSFA/SFA Exception Interception Co-authored-by: wangke293<wangke293@huawei.com> # message auto-generated for no-merge-commit merge: !3703 merge 7.3.0 into 7.3.0 QSFA/SFA Exception Interception Created-by: wangke293 Commit-by: wangke293 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: QSFA/SFA Exception Interception Special notes for your reviewers: See merge request: Ascend/op-plugin!3703	10 个月前
KvRmsNormRopeCacheNpuOpApi.cpp	[fix]修改v1分支下KvRmsNormRopeCache的输出shape Co-authored-by: zerosaki_admin<chenzhaominyi1@huawei.com> # message auto-generated for no-merge-commit merge: !4280 merge kv_730 into 7.3.0 [fix]修改v1分支下KvRmsNormRopeCache的输出shape Created-by: zerosaki_admin Commit-by: zerosaki_admin Merged-by: ascend-robot Description: <!-- PR描述模板更新日期：20260203 --> # 【合入来源】 > <font color="red">如有社区issue，请关联issue链接</font>\ > <font color="red">请勿携带内部流程信息（需求链接、问题单、内部issue等）</font> - [ ] 需求 - [x] 问题单 - [ ] issue/工单 - [ ] 重构优化 - [ ] 资料更新 # 【修改方案】修复KvRmsNormRopeCache算子在V1模式下的维度计算逻辑错误。原代码在非V2模式下，错误地使用了硬编码的固定尺寸来设置输出张量k_rope和c_kv的特征维度，导致当输入张量cos和gamma的实际维度与预设值不符时，算子内部计算会出错。本次修改旨在根据输入张量的实际维度进行动态计算，确保算子在不同输入配置下的兼容性和正确性。 # 【资料变更】不涉及 # 【接口变更】不涉及 # 【功能验证】验证通过 # 【CheckList】 - [x] 代码注释完备，正确记录错误日志 - [x] 代码实现进行了返回值、空指针等校验 - [x] PR标题正确使用类型标签，如：feat、fix、refactor、docs、test等 - [x] PR持续集成流水线（CI）执行通过，代码检查无异常 See merge request: Ascend/op-plugin!4280	8 个月前
LayerNormBackwardKernelNpuOpApi.cpp	!2275 fix cleancode Merge pull request !2275 from 詹昊/cleancode1	1 年前
LayerNormKernelNpuOpApi.cpp	!2326 [Task] cleancode 2/N. Merge pull request !2326 from will-devil/cleancode-1	1 年前
LeKernelNpuOpApi.cpp	!2272 Change dtype check for le. Merge pull request !2272 from 王夏夏/master	1 年前
LightningIndexerGradKernelNpuOpApi.cpp	add dsa Co-authored-by: DaiFu<daifu2@huawei.com> # message auto-generated for no-merge-commit merge: !3578 merge 7.3.0 into 7.3.0 add dsa Created-by: daifu1234567 Commit-by: DaiFu Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > /kind task What does this PR do / why do we need it: 新增LIG、SFAG及SLIGKLoss适配。 Special notes for your reviewers: 已本地单api和模型侧验证，无功能问题。用例及对外文档后续补充。 ![image.png](https://raw.gitcode.com/user-images/assets/7403085/9d809139-38cf-4e3b-b189-5db99637194b/image.png 'image.png') ![image.png](https://raw.gitcode.com/user-images/assets/7403085/1550bb18-edae-4ac5-8948-64061ab1a51d/image.png 'image.png') ![image.png](https://raw.gitcode.com/user-images/assets/7403085/b51a75ff-e530-4be1-9ec8-30fbd0cd960a/image.png 'image.png') See merge request: Ascend/op-plugin!3578	10 个月前
LightningIndexerKernelNpuOpApi.cpp	add dsa Co-authored-by: DaiFu<daifu2@huawei.com> # message auto-generated for no-merge-commit merge: !3578 merge 7.3.0 into 7.3.0 add dsa Created-by: daifu1234567 Commit-by: DaiFu Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > /kind task What does this PR do / why do we need it: 新增LIG、SFAG及SLIGKLoss适配。 Special notes for your reviewers: 已本地单api和模型侧验证，无功能问题。用例及对外文档后续补充。 ![image.png](https://raw.gitcode.com/user-images/assets/7403085/9d809139-38cf-4e3b-b189-5db99637194b/image.png 'image.png') ![image.png](https://raw.gitcode.com/user-images/assets/7403085/1550bb18-edae-4ac5-8948-64061ab1a51d/image.png 'image.png') ![image.png](https://raw.gitcode.com/user-images/assets/7403085/b51a75ff-e530-4be1-9ec8-30fbd0cd960a/image.png 'image.png') See merge request: Ascend/op-plugin!3578	10 个月前
LinalgNormKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
LinalgQrKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
LinalgSolveTriangularKernelNpuOpApi.cpp	!2313 modify cleancode/10 Merge pull request !2313 from 史佳馨/master	1 年前
LinearKernelNpuOpApi.cpp	support undocumented cube_math_type Co-authored-by: freezee<gengdongjie@huawei.com> # message auto-generated for no-merge-commit merge: !3834 merge 7.3.0 into 7.3.0 support undocumented cube_math_type Created-by: freezee Commit-by: freezee Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3834	9 个月前
LinspaceKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
LogKernelNpuOpApi.cpp	!2694 [Feature] Support named tensor [4/4]. Merge pull request !2694 from 刘嘉巍/named-dev	1 年前
LogSigmoidNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
LogSumExpKernelNpuOpApi.cpp	!2313 modify cleancode/10 Merge pull request !2313 from 史佳馨/master	1 年前
LogicalOrKernelNpuOpApi.cpp	!1666 [Feature] Codegen for multi-named tensor. Merge pull request !1666 from 刘嘉巍/master-add	1 年前
LogitBackwardKernelOpApi.cpp	!2587 fix logit_backward Merge pull request !2587 from shaoyf/logit_backward	1 年前
LogitKernelNpuOpApi.cpp	!2576 add logit and logit_backward Merge pull request !2576 from shaoyf/logit	1 年前
LtKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
MaskedFillKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
MaskedScatterKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
MaskedSelectKernelNpuOpApi.cpp	!2596 [DFX] More detail info in op_log and support common OP_LOG. Merge pull request !2596 from 刘嘉巍/master-dev	1 年前
MaskedSoftMaxKernelNpuOpApi.cpp	!2555 [Feature] Adapte _native_multi_head_attention. Merge pull request !2555 from 刘嘉巍/dev	1 年前
MatmulAllReduceBaseKernelNpuOpApi.cpp	!2235 cleancode Merge pull request !2235 from DaiFu/2503CleanCode	1 年前
MatmulBackwardKernelNpuOpApi.cpp	support undocumented cube_math_type Co-authored-by: freezee<gengdongjie@huawei.com> # message auto-generated for no-merge-commit merge: !3834 merge 7.3.0 into 7.3.0 support undocumented cube_math_type Created-by: freezee Commit-by: freezee Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3834	9 个月前
MatmulDoubleBackwardKernelNpuOpApi.cpp	support undocumented cube_math_type Co-authored-by: freezee<gengdongjie@huawei.com> # message auto-generated for no-merge-commit merge: !3834 merge 7.3.0 into 7.3.0 support undocumented cube_math_type Created-by: freezee Commit-by: freezee Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3834	9 个月前
MatmulKernelNpuOpApi.cpp	support undocumented cube_math_type Co-authored-by: freezee<gengdongjie@huawei.com> # message auto-generated for no-merge-commit merge: !3834 merge 7.3.0 into 7.3.0 support undocumented cube_math_type Created-by: freezee Commit-by: freezee Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3834	9 个月前
MatmulReduceScatterBaseKernelNpuOpApi.cpp	npu_mm_reduce_scatter_base support quant and comm_mode aiv Co-authored-by: wang-guangbin<wgb_strive@163.com> # message auto-generated for no-merge-commit merge: merge mm_rs into master npu_mm_reduce_scatter_base support quant and comm_mode aiv Created-by: wang-guangbin Commit-by: wang-guangbin Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: 本地验证通过 ![image.png](https://raw.gitcode.com/user-images/assets/7403085/3b23aa82-4d92-4ae0-9617-b48ea4f056e1/image.png 'image.png') See merge request: Ascend/op-plugin!3194	1 年前
MaxKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
MaxPool2dWithIndicesKernelNpuOpApi.cpp	!1666 [Feature] Codegen for multi-named tensor. Merge pull request !1666 from 刘嘉巍/master-add	1 年前
MaxPool3dWithIndicesBackwardKernelNpuOpApi.cpp	!1948 Add Adaption for MaxPool3d Forward and Backward Merge pull request !1948 from 柳宗谷/aclnnMaxPool3dWithArgmaxBackward	1 年前
MaxPool3dWithIndicesKernelNpuOpApi.cpp	!1948 Add Adaption for MaxPool3d Forward and Backward Merge pull request !1948 from 柳宗谷/aclnnMaxPool3dWithArgmaxBackward	1 年前
MeanKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
MinKernelNpuOpApi.cpp	!2235 cleancode Merge pull request !2235 from DaiFu/2503CleanCode	1 年前
MishKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
MlaPrologV2KernelNpuOpApi.cpp	!2689 delete line break of error message Merge pull request !2689 from vitamin123zsj/master	1 年前
MlaPrologV3KernelNpuOpApi.cpp	update cann version Co-authored-by: Le_666<yuanle2@huawei.com> # message auto-generated for no-merge-commit merge: !3914 merge 7.3.0 into 7.3.0 update cann version Created-by: Le_666 Commit-by: Le_666 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 兼容社区包alpha002版本 Special notes for your reviewers: See merge request: Ascend/op-plugin!3914	9 个月前
MmKernelNpuOpApi.cpp	support undocumented cube_math_type Co-authored-by: freezee<gengdongjie@huawei.com> # message auto-generated for no-merge-commit merge: !3834 merge 7.3.0 into 7.3.0 support undocumented cube_math_type Created-by: freezee Commit-by: freezee Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3834	9 个月前
MoeDistributeCombineAddRmsNormKernelOpApi.cpp	!3062 DispatchV3/CombineV3/CombineaddrmsnormV2 update Merge pull request !3062 from Oblivion/master	1 年前
MoeDistributeCombineKernelOpApi.cpp	!2838 revert change of moe_distribute_dispatch/combine v1 interface Merge pull request !2838 from sincatter/master_dis_v1	1 年前
MoeDistributeCombineKernelV2OpApi.cpp	适配dispatch\combine性能打点 Co-authored-by: lyt_claire<luyitong1@huawei.com> # message auto-generated for no-merge-commit merge: !3793 merge performanceFix730 into 7.3.0 适配dispatch\combine性能打点 Created-by: lyt_claire Commit-by: lyt_claire Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? What does this PR do / why do we need it: PTA适配dispatch\combine性能打点，修复v3接口performance_info传空tensor拦截问题 Special notes for your reviewers: See merge request: Ascend/op-plugin!3793	10 个月前
MoeDistributeDispatchKernelOpApi.cpp	!2838 revert change of moe_distribute_dispatch/combine v1 interface Merge pull request !2838 from sincatter/master_dis_v1	1 年前
MoeDistributeDispatchV2KernelOpApi.cpp	适配dispatch\combine性能打点 Co-authored-by: lyt_claire<luyitong1@huawei.com> # message auto-generated for no-merge-commit merge: !3793 merge performanceFix730 into 7.3.0 适配dispatch\combine性能打点 Created-by: lyt_claire Commit-by: lyt_claire Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? What does this PR do / why do we need it: PTA适配dispatch\combine性能打点，修复v3接口performance_info传空tensor拦截问题 Special notes for your reviewers: See merge request: Ascend/op-plugin!3793	10 个月前
MoeFinalizeRoutingV2KernelNpuOpApi.cpp	!2488 Refactor compability judgement. Merge pull request !2488 from 刘嘉巍/master	1 年前
MoeGatingTopKKernelNpuOpApi.cpp	!2377 add npu_moe_gating_top_k api Merge pull request !2377 from 汤PC/master	1 年前
MoeGatingTopKSoftmaxKernelNpuOpApi.cpp	!2233 Fix cleancode. Merge pull request !2233 from 王夏夏/master	1 年前
MoeInitRoutingKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
MoeInitRoutingQuantKernelNpuOpApi.cpp	!2930 新增MoeInitRoutingQuant算子接口 Merge pull request !2930 from CY-Slightwind/master_cy	1 年前
MoeInitRoutingV2KernelNpuOpApi.cpp	special case divert to moe init routing V2 Co-authored-by: yaolu_happy<yaolu15@huawei.com> # message auto-generated for no-merge-commit merge: merge moe_init_routing_v3 into master special case divert to moe init routing V2 Created-by: yaolu_happy Commit-by: yaolu_happy Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3227	1 年前
MoeReRoutingOpApi.cpp	!2633 sync A5 Merge pull request !2633 from 宋恺/master	1 年前
MoeTokenPermuteWithRoutingMapGradOpApi.cpp	!2871 Add npu_moe_token_permute_with_routing_map and testcase Merge pull request !2871 from wgb/bugfix	1 年前
MoeTokenPermuteWithRoutingMapOpApi.cpp	!2871 Add npu_moe_token_permute_with_routing_map and testcase Merge pull request !2871 from wgb/bugfix	1 年前
MoeTokenUnpermuteWithRoutingMapGradKernelNpuOpApi.cpp	!2727 [Feature] npu_moe_token_unpermute_with_routing_map Merge pull request !2727 from 刘嘉巍/mha	1 年前
MoeTokenUnpermuteWithRoutingMapKernelOpApi.cpp	!2727 [Feature] npu_moe_token_unpermute_with_routing_map Merge pull request !2727 from 刘嘉巍/mha	1 年前
MseLossKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
MulKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
MultilabelMarginLossKernelNpuOpApi.cpp	!2295 [3/N]cleancode Merge pull request !2295 from zhangqiongwen/master_cleancode_0318_01	1 年前
MultinomialKernelNpuOpApi.cpp	!3078 support generator op with capture Merge pull request !3078 from 闫鹏全/master	1 年前
MvKernelNpuOpApi.cpp	!2326 [Task] cleancode 2/N. Merge pull request !2326 from will-devil/cleancode-1	1 年前
NLLLoss2dBackwardKernelNpuOpApi.cpp	!2275 fix cleancode Merge pull request !2275 from 詹昊/cleancode1	1 年前
NLLLoss2dKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
NLLLossBackwardKernelNpuOpApi.cpp	!2316 cleancode Merge pull request !2316 from DaiFu/2503CleanCode1	1 年前
NLLLossKernelNpuOpApi.cpp	!2246 [Codecheck] Fix codecheck. Merge pull request !2246 from 刘嘉巍/deformable	1 年前
NanToNumKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
NansumKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
NativeDropoutKernelNpuOpApi.cpp	!3134 Optimize the operator adaptation of dropout in the aclgraph scenario. Merge pull request !3134 from 闫鹏全/master_dropout_new.py	1 年前
NativeMultiHeadAttentionKernelNpuOpApi.cpp	!2555 [Feature] Adapte _native_multi_head_attention. Merge pull request !2555 from 刘嘉巍/dev	1 年前
NeKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
NnpackSpatialConvolutionKernelNpuOpApi.cpp	!3099 [Feature] aclnn for _nnpack_spatial_convolution aclnn Merge pull request !3099 from 刘嘉巍/master-dev	1 年前
NonzeroKernelNpuOpApi.cpp	!2596 [DFX] More detail info in op_log and support common OP_LOG. Merge pull request !2596 from 刘嘉巍/master-dev	1 年前
NormKernelNpuOpApi.cpp	!2303 Norm fix with inf or -inf Merge pull request !2303 from 王超/norm_fix	1 年前
NormalKernelNpuOpApi.cpp	!3078 support generator op with capture Merge pull request !3078 from 闫鹏全/master	1 年前
ObfuscationCalculateKernelNpuOpApi.cpp	!3043 Add obfuscation_initialize obfucation_finialize obfuscation_calculate Merge pull request !3043 from 蒋学博/master	1 年前
ObfuscationFinalizeKernelNpuOpApi.cpp	!3043 Add obfuscation_initialize obfucation_finialize obfuscation_calculate Merge pull request !3043 from 蒋学博/master	1 年前
ObfuscationInitializeKernelNpuOpApi.cpp	!3043 Add obfuscation_initialize obfucation_finialize obfuscation_calculate Merge pull request !3043 from 蒋学博/master	1 年前
OneHotKernelNpuOpApi.cpp	fix onehot error report for dtensor Co-authored-by: gcw_Dgfy2aKk<dujunjie2@huawei.com> # message auto-generated for no-merge-commit merge: !3906 merge fix_onthot_dtensor_730 into 7.3.0 fix onehot error report for dtensor Created-by: gcw_Dgfy2aKk Commit-by: gcw_Dgfy2aKk Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: F.one_hot(dtensor) used to run into a mixed DTensor-Tensor operation due to an arange call creating a new Tensor (not DTensor). This PR fixes it by allowing implicit replication of Tensors for the arange call and the one consumer of the arange call (the at::eq call). Special notes for your reviewers: torch version ≥ 2.9.0 See merge request: Ascend/op-plugin!3906	9 个月前
OnesKernelNpuOpApi.cpp	!2309 fix cleancode Merge pull request !2309 from 詹昊/master	1 年前
OnesLikeKernelNpuOpApi.cpp	Fix the torch.ones_like adaptation on the NPU side. Co-authored-by: shi-jiaxin9<shijiaxin10@h-partners.com> # message auto-generated for no-merge-commit merge: !3769 merge 7.3.0 into 7.3.0 Fix the torch.ones_like adaptation on the NPU side. Created-by: shi-jiaxin9 Commit-by: shi-jiaxin9 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug What does this PR do / why do we need it: Fix the torch.ones_like adaptation on the NPU side. Special notes for your reviewers: See merge request: Ascend/op-plugin!3769	10 个月前
PackPaddedSequenceKernelNpuOpApi.cpp	!3098 Synchronizes selected interfaces from aclop to aclnn Merge pull request !3098 from shaoyf/aclop_3	1 年前
PadPackedSequenceKernelNpuOpApi.cpp	!3098 Synchronizes selected interfaces from aclop to aclnn Merge pull request !3098 from shaoyf/aclop_3	1 年前
PdistKernelNpuOpApi.cpp	!2247 cleancode for ops. Merge pull request !2247 from 王夏夏/master	1 年前
PrefetchKernelNpuOpApi.cpp	!2313 modify cleancode/10 Merge pull request !2313 from 史佳馨/master	1 年前
PreluBackwardKernelNpuOpApi.cpp	!2313 modify cleancode/10 Merge pull request !2313 from 史佳馨/master	1 年前
ProdKernelNpuOpApi.cpp	!2322 modify cleancode/3_1 Merge pull request !2322 from 史佳馨/master	1 年前
PutKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
QuantLightningIndexerKernelNpuOpApi.cpp	新增算子LI\QLI\SFA\QSFA Co-authored-by: zzzyh22<zhangyaohui22@huawei.com> # message auto-generated for no-merge-commit merge: !3521 merge 7.3.0 into 7.3.0 新增算子LI\QLI\SFA\QSFA Created-by: zzzyh22 Commit-by: zzzyh22 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > /kind bug > /kind task > >kind feature What does this PR do / why do we need it: 根据需求新增LI\QLI\SFA\QSFA四个算子，其中 SFA算子通过“只计算关键部分”大幅减少计算量，优化算子性能。 QSFA在SFA的基础上支持了Per-Token-Head-Tile-128量化输入。 LI算子基于一系列操作得到每一个 token 对应的 Top-k个位置。 LIQ在LI的基础上支持了Per-Token-Head量化输入。 Special notes for your reviewers: See merge request: Ascend/op-plugin!3521	10 个月前
QuantMatmulKernelNpuOpApi.cpp	[master] aclnnQuantMatmulV5 W4A4 support for PerGroup quantization Co-authored-by: Slightwind<caoyi31@huawei.com> # message auto-generated for no-merge-commit merge: !3314 merge master_w4a4_pergroup into master [master] aclnnQuantMatmulV5 W4A4 support for PerGroup quantization Created-by: Slightwind Commit-by: Slightwind Merged-by: ascend-robot Description: * 新增了`npu_quant_matmul`接口对W4A4 PerGroup场景的支持，底层调用aclnnQuantMatmulV5算子。 See merge request: Ascend/op-plugin!3314	1 年前
QuantMatmulReduceSumKernelNpuOpApi.cpp	!2806 add npu_quant_matmul_reduce_sum Merge pull request !2806 from AaronTang/qbmm_reduce_sum	1 年前
QuantScatterKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
QuantizeKernelNpuOpApi.cpp	npu_quantize接口input数据格式支持NZ Co-authored-by: wkp<1289706727@qq.com> # message auto-generated for no-merge-commit merge: merge master-quantize into master npu_quantize接口input数据格式支持NZ Created-by: kunpengW Commit-by: wkp Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: ![image.png](https://raw.gitcode.com/user-images/assets/7403085/0984ff44-7fd0-4bd0-aac1-f621b7a6a750/image.png 'image.png') See merge request: Ascend/op-plugin!3219	1 年前
QuantizePerChannelKernelNpuOpApi.cpp	!3079 [Feature] aclnn for several apis Merge pull request !3079 from 刘嘉巍/meta	1 年前
QuantizePerTensorKernelNpuOpApi.cpp	!3079 [Feature] aclnn for several apis Merge pull request !3079 from 刘嘉巍/meta	1 年前
RReluWithNoiseKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
RandomKernelNpuOpApi.cpp	!3078 support generator op with capture Merge pull request !3078 from 闫鹏全/master	1 年前
RandpermKernelNpuOpApi.cpp	!2282 Fix codecheck. Merge pull request !2282 from 刘嘉巍/cleancode	1 年前
RangeKernelNpuOpApi.cpp	!1971 主线，对标竞品适配range/arange op-plugin Merge pull request !1971 from 王振动/mst_range_1105	1 年前
RecurrentGatedDeltaRuleKernelNpuOpApi.cpp	Add recurrent_gated_delta_rule Co-authored-by: NeverLCY<liuchunyuan@huawei.com> # message auto-generated for no-merge-commit merge: !3602 merge RGDR_7.3.0 into 7.3.0 Add recurrent_gated_delta_rule Created-by: NeverLCY Commit-by: NeverLCY Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3602	10 个月前
RemainderKernelNpuOpApi.cpp	!2245 [Codecheck] Fix codecheck. Merge pull request !2245 from 刘嘉巍/x1	1 年前
RenormKernelNpuOpApi.cpp	!2260 fix cleancode Merge pull request !2260 from 詹昊/master	1 年前
RepeatInterLeaveKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
RepeatInterleaveBackwardKernelOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
RmsNormBackwardKernelOpApi.cpp	!2247 cleancode for ops. Merge pull request !2247 from 王夏夏/master	1 年前
RmsNormKernelOpApi.cpp	!2247 cleancode for ops. Merge pull request !2247 from 王夏夏/master	1 年前
RoiAlignBackwardKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
RoiAlignKernelNpuOpApi.cpp	!2313 modify cleancode/10 Merge pull request !2313 from 史佳馨/master	1 年前
RopeQuantKVCacheKernelNpuOpApi.cpp	!2103 modify op dequant_repo_quant_kv_cache file name Merge pull request !2103 from 流浪者LingLong/master	1 年前
RotaryMulBackwardKernelNpuOpApi.cpp	!2790 Revert some files Merge pull request !2790 from chuboning/master	1 年前
RotaryMulKernelNpuOpApi.cpp	!2790 Revert some files Merge pull request !2790 from chuboning/master	1 年前
RoundDecimalsKernelNpuOpApi.cpp	Change the error message to match aclnnRound Co-authored-by: qq_43150381<zhugelu1@h-partners.com> # message auto-generated for no-merge-commit merge: !3688 merge 7.3.0 into 7.3.0 Change the error message to match aclnnRound Created-by: qq_43150381 Commit-by: qq_43150381 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3688	10 个月前
ScaledDotProductAttentionKernelNpuOpApi.cpp	sdpa for graph Co-authored-by: DaiFu<daifu2@huawei.com> # message auto-generated for no-merge-commit merge: !3420 merge 2510sdpa_graph into master sdpa for graph Created-by: daifu1234567 Commit-by: DaiFu Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind feature What does this PR do / why do we need it: Support SDPA for graph mode. 1. op_api命名空间会直接走进aclnn的实现，导致问题，因此改成at重新分发 2. empty_like会保持非连续性，而eager模式下用apply_tensor_without_format申请的tensor是连续的，导致stride对不上，因此需要将empty_like改成new_empty以申请连续tensor Special notes for your reviewers: See merge request: Ascend/op-plugin!3420	11 个月前
ScatterAddKernelNpuOpApi.cpp	!2206 [Codecheck] Fix codecheck. Merge pull request !2206 from 刘嘉巍/deformable	1 年前
ScatterKernelNpuOpApi.cpp	!2293 [2/N] cleancode Merge pull request !2293 from zhangqiongwen/master_cleancode_0317	1 年前
ScatterListKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
ScatterNdUpdateKernelNpuOpApi.cpp	!2353 [Task] cleancode 3/N. Merge pull request !2353 from will-devil/cleancode-2	1 年前
ScatterPaKvCacheNpuOpApi.cpp	fix scatter_pa_kv_cache input Co-authored-by: wy519<wangyang519@huawei.com> # message auto-generated for no-merge-commit merge: !3252 merge scatter_pa_kv_cache_master into master fix scatter_pa_kv_cache input Created-by: wy519 Commit-by: wy519 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > /kind feature Interface Adaptation Operator IR Changes What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3252	1 年前
ScatterUpdateKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
SignKernelNpuOpApi.cpp	!2694 [Feature] Support named tensor [4/4]. Merge pull request !2694 from 刘嘉巍/named-dev	1 年前
SilentCheckV2KernelNpuOpApi.cpp	!2314 modify cleancode/11 Merge pull request !2314 from 史佳馨/master	1 年前
SilentCheckV3KernelNpuOpApi.cpp	!2314 modify cleancode/11 Merge pull request !2314 from 史佳馨/master	1 年前
SimExponentialNpuOpApi.cpp	fix increment Co-authored-by: liangsongwei<liangsongwei@huawei.com> # message auto-generated for no-merge-commit merge: !3627 merge 7.3.0 into 7.3.0 fix increment Created-by: liangsongwei Commit-by: liangsongwei Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3627	10 个月前
SlogdetKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
SmoothL1LossBackwrdKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
SmoothL1LossKernelNpuOpApi.cpp	!2640 fix smoothl1loss blending accuracy Merge pull request !2640 from liuxiaoxing/dev_david	1 年前
SoftmaxKernelNpuOpApi.cpp	!3096 修改不合理的内存分配问题 Merge pull request !3096 from qianlihai/master	1 年前
SortKernelNpuOpApi.cpp	!1713 fix ops * fix ops	2 年前
SparseFlashAttentionGradKernelNpuOpApi.cpp	add dsa Co-authored-by: DaiFu<daifu2@huawei.com> # message auto-generated for no-merge-commit merge: !3578 merge 7.3.0 into 7.3.0 add dsa Created-by: daifu1234567 Commit-by: DaiFu Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > /kind task What does this PR do / why do we need it: 新增LIG、SFAG及SLIGKLoss适配。 Special notes for your reviewers: 已本地单api和模型侧验证，无功能问题。用例及对外文档后续补充。 ![image.png](https://raw.gitcode.com/user-images/assets/7403085/9d809139-38cf-4e3b-b189-5db99637194b/image.png 'image.png') ![image.png](https://raw.gitcode.com/user-images/assets/7403085/1550bb18-edae-4ac5-8948-64061ab1a51d/image.png 'image.png') ![image.png](https://raw.gitcode.com/user-images/assets/7403085/b51a75ff-e530-4be1-9ec8-30fbd0cd960a/image.png 'image.png') See merge request: Ascend/op-plugin!3578	10 个月前
SparseFlashAttentionKernelNpuOpApi.cpp	QSFA/SFA Exception Interception Co-authored-by: wangke293<wangke293@huawei.com> # message auto-generated for no-merge-commit merge: !3703 merge 7.3.0 into 7.3.0 QSFA/SFA Exception Interception Created-by: wangke293 Commit-by: wangke293 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: QSFA/SFA Exception Interception Special notes for your reviewers: See merge request: Ascend/op-plugin!3703	10 个月前
SparseLightningIndexerGradKLLossKernelNpuOpApi.cpp	add dsa Co-authored-by: DaiFu<daifu2@huawei.com> # message auto-generated for no-merge-commit merge: !3578 merge 7.3.0 into 7.3.0 add dsa Created-by: daifu1234567 Commit-by: DaiFu Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > /kind task What does this PR do / why do we need it: 新增LIG、SFAG及SLIGKLoss适配。 Special notes for your reviewers: 已本地单api和模型侧验证，无功能问题。用例及对外文档后续补充。 ![image.png](https://raw.gitcode.com/user-images/assets/7403085/9d809139-38cf-4e3b-b189-5db99637194b/image.png 'image.png') ![image.png](https://raw.gitcode.com/user-images/assets/7403085/1550bb18-edae-4ac5-8948-64061ab1a51d/image.png 'image.png') ![image.png](https://raw.gitcode.com/user-images/assets/7403085/b51a75ff-e530-4be1-9ec8-30fbd0cd960a/image.png 'image.png') See merge request: Ascend/op-plugin!3578	10 个月前
SplitWithSizesCopyNpuOpApi.cpp	[fix]split_with_sizes_copy performance reduction Co-authored-by: zqwenn<zhangqiongwen@huawei.com> # message auto-generated for no-merge-commit merge: !3494 merge 7.3.0_split_fix into 7.3.0 [fix]split_with_sizes_copy performance reduction Created-by: zqwenn Commit-by: zqwenn Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 融合算子只在部分场景性能有优化，非优化场景下aclnn的性能劣于小算子拼接性能。这个pr修复性能出现劣化的场景，只保留特定场景接入aclnn，其余场景走小算子。 Special notes for your reviewers: See merge request: Ascend/op-plugin!3494	11 个月前
StdKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
StdMeanKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
StftKernelNpuOpApi.cpp	!2260 fix cleancode Merge pull request !2260 from 詹昊/master	1 年前
SubKernelNpuOpApi.cpp	!2317 Cleancode for ops. Merge pull request !2317 from 王夏夏/cleancode	1 年前
SumKernelNpuOpApi.cpp	!3132 Delete code for v1.11.0 and v2.0 (7/n) Merge pull request !3132 from 刘嘉巍/master-dev	1 年前
SwigluquantKernelNpuOpApi.cpp	add torch_npu.npu_swiglu_quant api, based on aclnnSwiGluQuantV2 Co-authored-by: zhaoyu20121997<nanzhaogang@qq.com> # message auto-generated for no-merge-commit merge: merge swiglu_pr into master add torch_npu.npu_swiglu_quant api, based on aclnnSwiGluQuantV2 Created-by: zhaoyu20121997 Commit-by: zhaoyu20121997 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > /kind feature What does this PR do / why do we need it: 新增torch_npu.npu_swiglu_quant接口，接入aclnnSwiGluQuantV2算子 Special notes for your reviewers: 接口说明请参考提交的文件docs/context/torch_npu-npu_swiglu_quant.md See merge request: Ascend/op-plugin!3167	1 年前
TanhKernelNpuOpApi.cpp	!2694 [Feature] Support named tensor [4/4]. Merge pull request !2694 from 刘嘉巍/named-dev	1 年前
ThresholdKernelNpuOpApi.cpp	!2694 [Feature] Support named tensor [4/4]. Merge pull request !2694 from 刘嘉巍/named-dev	1 年前
TopKKernelNpuOpApi.cpp	!2204 clean code : infershape file Merge pull request !2204 from wgb/clean_code	1 年前
TransQuantParamKernelNpuOpApi.cpp	!2488 Refactor compability judgement. Merge pull request !2488 from 刘嘉巍/master	1 年前
TransfromBiasRescaleQKVKernelNpuOpApi.cpp	!2555 [Feature] Adapte _native_multi_head_attention. Merge pull request !2555 from 刘嘉巍/dev	1 年前
UnfoldBackwardLernelNpuOpApi.cpp	[fix] update unfold_backward check cann version Co-authored-by: zqwenn<zhangqiongwen@huawei.com> # message auto-generated for no-merge-commit merge: merge master_unfold_backward_cann_check into master [fix] update unfold_backward check cann version Created-by: zqwenn Commit-by: zqwenn Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: Special notes for your reviewers: See merge request: Ascend/op-plugin!3215	1 年前
UniformKernelNpuOpApi.cpp	!3078 support generator op with capture Merge pull request !3078 from 闫鹏全/master	1 年前
Unique2KernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UniqueConsecutiveKernelNpuOpApi.cpp	solve the problem of high memory usage in tensor.unique_consecutive Co-authored-by: zhaoyu20121997<nanzhaogang@qq.com> # message auto-generated for no-merge-commit merge: merge uniqueConsecutive into master solve the problem of high memory usage in tensor.unique_consecutive Created-by: zhaoyu20121997 Commit-by: zhaoyu20121997 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug What does this PR do / why do we need it: 解决tensor.unique_consecutive接口显存占用过高的问题： Special notes for your reviewers: 之前的逻辑：当传入的dim不为空时就为y_inverse和y_counts申请显存；修改后的逻辑：当return_inverse、return_counts为True时，才为y_inverse和y_counts申请显存，否则为其赋值空tensor，避免申请大量无意义的显存； See merge request: Ascend/op-plugin!3184	1 年前
UniqueDimKernelNpuOpApi.cpp	!2596 [DFX] More detail info in op_log and support common OP_LOG. Merge pull request !2596 from 刘嘉巍/master-dev	1 年前
UnsafeIndexKernelNpuOpApi.cpp	!3086 The aclop for torch._dim_arange, torch._unsafe_index, torch.index_add, and torch.nn.functional.adaptive_avg_pool1d has been switched to aclnn. Merge pull request !3086 from 史佳馨/master	1 年前
UpsampleBicubic2dAABackwardKernelNpuOpApi.cpp	!2213 clean code Merge pull request !2213 from wgb/clean_code1	1 年前
UpsampleBicubic2dAAKernelNpuOpApi.cpp	!2275 fix cleancode Merge pull request !2275 from 詹昊/cleancode1	1 年前
UpsampleBicubic2dBackwardKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UpsampleBicubic2dKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UpsampleBilinear2dAABackwardKernelNpuOpApi.cpp	!2213 clean code Merge pull request !2213 from wgb/clean_code1	1 年前
UpsampleBilinear2dAAKernelNpuOpApi.cpp	!2204 clean code : infershape file Merge pull request !2204 from wgb/clean_code	1 年前
UpsampleBilinear2dBackwardKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UpsampleBilinear2dKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UpsampleLinear1dBackwardKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UpsampleLinear1dKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UpsampleNearest1dBackwardKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UpsampleNearest1dKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UpsampleNearest2dBackwardKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UpsampleNearest2dKernelOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UpsampleNearest3dBackwardKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UpsampleNearest3dKernelOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UpsampleNearestExact1dBackwardKernelNpuOpApi.cpp	!2314 modify cleancode/11 Merge pull request !2314 from 史佳馨/master	1 年前
UpsampleNearestExact1dKernelOpApi.cpp	!2276 fix cleancode Merge pull request !2276 from 詹昊/cleancode2	1 年前
UpsampleNearestExact2dBackwardKernelNpuOpApi.cpp	!2199 NearestExact兼容性修改 Merge pull request !2199 from 陈海杰/master	1 年前
UpsampleNearestExact2dKernelOpApi.cpp	!2199 NearestExact兼容性修改 Merge pull request !2199 from 陈海杰/master	1 年前
UpsampleNearestExact3dBackwardKernelOpApi.cpp	!2199 NearestExact兼容性修改 Merge pull request !2199 from 陈海杰/master	1 年前
UpsampleNearestExact3dKernelOpApi.cpp	!2353 [Task] cleancode 3/N. Merge pull request !2353 from will-devil/cleancode-2	1 年前
UpsampleTrilinear3dBackwardKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
UpsampleTrilinear3dKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
VarKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
ViesAsRealKernelNpuOpApi.cpp	!3098 Synchronizes selected interfaces from aclop to aclnn Merge pull request !3098 from shaoyf/aclop_3	1 年前
WeightNormKernelNpuOpApi.cpp	!3098 Synchronizes selected interfaces from aclop to aclnn Merge pull request !3098 from shaoyf/aclop_3	1 年前
WeightQuantBatchMatmulV2KernelNpuOpApi.cpp	!2285 Fix codecheck. Merge pull request !2285 from 刘嘉巍/x1	1 年前
WhereKernelNpuOpApi.cpp	fix where for scalar tensor Co-authored-by: DaiFu<daifu2@huawei.com> # message auto-generated for no-merge-commit merge: !3735 merge 7.3.0 into 7.3.0 fix where for scalar tensor Created-by: daifu1234567 Commit-by: DaiFu Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug What does this PR do / why do we need it: fix where for scalar tensor Special notes for your reviewers: See merge request: Ascend/op-plugin!3735	10 个月前
ZerosKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (6/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge new into master Delete code for v1.11.0 and v2.0 (6/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3186	1 年前
ZerosLikeKernelNpuOpApi.cpp	!2265 fix cleancode Merge pull request !2265 from 詹昊/cleancode1	1 年前
_UniqueKernelNpuOpApi.cpp	Delete code for v1.11.0 and v2.0 (5/n) Co-authored-by: 刘嘉巍<843972097@qq.com> # message auto-generated for no-merge-commit merge: merge dev into master Delete code for v1.11.0 and v2.0 (5/n) Created-by: liu-jiaweikf Commit-by: 刘嘉巍 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> What type of PR is this? > Uncomment only one `/kind <>` line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature What does this PR do / why do we need it: 删除代码中仅支持v1.11/v2.0的部分 Special notes for your reviewers: See merge request: Ascend/op-plugin!3187	1 年前
__Ior__KernelNpuOpApi.cpp	!2316 cleancode Merge pull request !2316 from DaiFu/2503CleanCode1	1 年前
_thnnFusedFruCellNpuOpApi.cpp	!2759 Add _thnn_fused_gru_cell for pt2.5+ Merge pull request !2759 from shaoyf/master_gru_cell	1 年前