dev

ascend-robot[Feature][ops]aclnn编译工程适配

b748b834创建于 3月7日历史提交

文件	最后提交记录	最后更新时间
adalayernorm.cpp	[feature]新增adaLayernormv2的plugin Co-authored-by: chensiyu47<chensiyu49@huawei.com> # message auto-generated for no-merge-commit merge: !172 merge dev into dev [feature]新增adaLayernormv2的plugin Created-by: chensiyu47 Commit-by: chensiyu47 Merged-by: ascend-robot Description: <!-- PR描述模板更新日期：20251224 --> # Purpose 新增adaLayernormv2的plugin # Test Plan 单算子测试 # Test Report  See merge request: Ascend/MindIE-SD!172	8 个月前
adalayernorm.h	[feature]新增adaLayernormv2的plugin Co-authored-by: chensiyu47<chensiyu49@huawei.com> # message auto-generated for no-merge-commit merge: !172 merge dev into dev [feature]新增adaLayernormv2的plugin Created-by: chensiyu47 Commit-by: chensiyu47 Merged-by: ascend-robot Description: <!-- PR描述模板更新日期：20251224 --> # Purpose 新增adaLayernormv2的plugin # Test Plan 单算子测试 # Test Report  See merge request: Ascend/MindIE-SD!172	8 个月前
block_sparse_attention.cpp	[Feature][ops]aclnn编译工程适配 Co-authored-by: guowenna1<guowenna1@huawei.com> # message auto-generated for no-merge-commit merge: !201 merge aclnn_compil into dev [Feature][ops]aclnn编译工程适配 Created-by: guowenna1 Commit-by: guowenna1 Merged-by: ascend-robot Description: Purpose 1、算子接入方式从aclop改为aclnn，aclnn编译工程适配 2、移除rope的plugin定义，对外接口保留，但路由到torch npu实现 Test Plan A2、A3、A5编译工程测试 A2、A3全量测试用例验证 A2、A3模型整网验证 Test Report  See merge request: Ascend/MindIE-SD!201	7 个月前
block_sparse_attention.h	[feature]新增BlockSparseAttention算子的plugin和UT Co-authored-by: lanwangli<lanwangli@huawei.com> # message auto-generated for no-merge-commit merge: !73 merge bsa_op2_host into dev [feature]新增BlockSparseAttention算子的plugin和UT Created-by: mazhixin00_00 Commit-by: lanwangli Merged-by: ascend-robot Description: [feature]新增BlockSparseAttention算子、以及plugin和UT **单算子调用** import torch torch.ops.load_library("mindiesd/plugin/libPTAExtensionOPS.so") output = torch.ops.mindie.block_sparse_attention( query=query, key=key, value=value, sparse_mask=smask, sparse_count_table=sparse_count_table, input_layout="BNSD", num_heads=query.shape[1] , num_key_value_heads=key.shape[1], scale_value = query.shape[-1] **-0.5, causal=False ) - sparse_mask：稀疏 mask，shape 为 （BatchSize, NumHeads, Sn1, Sn2），其中 Sn1 = (SeqLenQ + SparseSize - 1)//SparseSize，Sn2 = Align((SeqLenKV + SparseSize - 1)//SparseSize, 32)，数据类型支持int8。 - sparse_count_table：为 sparse_mask 对最后一维的求和，shape 为 （BatchSize, NumHeads, Sn1），数据类型支持int32。 - input_layout：目前只支持BNSD、BSND和BSH **全量测试结果**  **单算子测试结果**  See merge request: Ascend/MindIE-SD!73	10 个月前
find_op_path.cpp	aclnn算子op plugin 公共接口 Co-authored-by: changetheway<guotaoyuan1@h-partners.com> # message auto-generated for no-merge-commit merge: !30 merge dev_1203_pta into dev aclnn算子op plugin 公共接口 Created-by: changetheway Commit-by: changetheway Merged-by: ascend-robot Description: aclnn算子op plugin 公共接口  See merge request: Ascend/MindIE-SD!30	10 个月前
find_op_path.h	aclnn算子op plugin 公共接口 Co-authored-by: changetheway<guotaoyuan1@h-partners.com> # message auto-generated for no-merge-commit merge: !30 merge dev_1203_pta into dev aclnn算子op plugin 公共接口 Created-by: changetheway Commit-by: changetheway Merged-by: ascend-robot Description: aclnn算子op plugin 公共接口  See merge request: Ascend/MindIE-SD!30	10 个月前
la.cpp	[Feature][ops]aclnn编译工程适配 Co-authored-by: guowenna1<guowenna1@huawei.com> # message auto-generated for no-merge-commit merge: !201 merge aclnn_compil into dev [Feature][ops]aclnn编译工程适配 Created-by: guowenna1 Commit-by: guowenna1 Merged-by: ascend-robot Description: Purpose 1、算子接入方式从aclop改为aclnn，aclnn编译工程适配 2、移除rope的plugin定义，对外接口保留，但路由到torch npu实现 Test Plan A2、A3、A5编译工程测试 A2、A3全量测试用例验证 A2、A3模型整网验证 Test Report  See merge request: Ascend/MindIE-SD!201	7 个月前
la.h	[dev]同步最新代码 Co-authored-by: mazhixin00_00<mazhixin7@huawei.com>	10 个月前
la_preprocess.cpp	[Feature][ops]aclnn编译工程适配 Co-authored-by: guowenna1<guowenna1@huawei.com> # message auto-generated for no-merge-commit merge: !201 merge aclnn_compil into dev [Feature][ops]aclnn编译工程适配 Created-by: guowenna1 Commit-by: guowenna1 Merged-by: ascend-robot Description: Purpose 1、算子接入方式从aclop改为aclnn，aclnn编译工程适配 2、移除rope的plugin定义，对外接口保留，但路由到torch npu实现 Test Plan A2、A3、A5编译工程测试 A2、A3全量测试用例验证 A2、A3模型整网验证 Test Report  See merge request: Ascend/MindIE-SD!201	7 个月前
la_preprocess.h	[dev]同步最新代码 Co-authored-by: mazhixin00_00<mazhixin7@huawei.com>	10 个月前
layernorm.cpp	[feature]新增aclnnLayerNorm的plugin和对外接口 Co-authored-by: mazhixin00_00<mazhixin7@huawei.com> # message auto-generated for no-merge-commit merge: !159 merge dev_layernorm into dev [feature]新增aclnnLayerNorm的plugin和对外接口 Created-by: mazhixin00_00 Commit-by: mazhixin00_00 Merged-by: ascend-robot Description: <!-- PR描述模板更新日期：20251224 --> # Purpose 新增aclnnLayerNorm的plugin # Test Plan 单算子测试 全量测试 # Test Report 单算子   See merge request: Ascend/MindIE-SD!159	8 个月前
layernorm.h	[feature]新增aclnnLayerNorm的plugin和对外接口 Co-authored-by: mazhixin00_00<mazhixin7@huawei.com> # message auto-generated for no-merge-commit merge: !159 merge dev_layernorm into dev [feature]新增aclnnLayerNorm的plugin和对外接口 Created-by: mazhixin00_00 Commit-by: mazhixin00_00 Merged-by: ascend-robot Description: <!-- PR描述模板更新日期：20251224 --> # Purpose 新增aclnnLayerNorm的plugin # Test Plan 单算子测试 全量测试 # Test Report 单算子   See merge request: Ascend/MindIE-SD!159	8 个月前
pytorch_npu_helper.h	[bugfix]修复注册多个aclnn算子时编译报错的问题 Co-authored-by: mazhixin00_00<mazhixin7@huawei.com> # message auto-generated for no-merge-commit merge: !53 merge dev into dev [bugfix]修复注册多个aclnn算子时编译报错的问题 Created-by: mazhixin00_00 Commit-by: mazhixin00_00 Merged-by: ascend-robot Description: [bugfix]修复注册多个aclnn算子时编译报错的问题 报错截图：  修改后成功编译  See merge request: Ascend/MindIE-SD!53	10 个月前
rainfusionattention.cpp	[feature]rf_v2算子支持BNSD Co-authored-by: mazhixin00_00<mazhixin7@huawei.com> # message auto-generated for no-merge-commit merge: !126 merge dev into dev [feature]rf_v2算子支持BNSD Created-by: mazhixin00_00 Commit-by: mazhixin00_00 Merged-by: ascend-robot Description: <!-- PR描述模板更新日期：20251224 --> # Which issue(s) this PR fixes or accomplishes 扩展了rf_v2算子对输入张量布局的支持范围，移除了原有sparse_type为rf_v2时input_layout参数仅支持"BSND"的限制。通过底层算子与前端逻辑的协同修改，现同时支持"BSND"和"BNSD"两种布局，提升了算子的通用性和灵活性，并解决了相关池化、掩码计算及张量重排操作在不同布局下的兼容性问题。 # Purpose rf_v2算子支持BNSD # Test Plan layout设置为bnsd，测试rfv2算子的功能和精度 # Test Report  See merge request: Ascend/MindIE-SD!126	9 个月前
rainfusionattention.h	[feature]新增RainFusionAttention算子的plugin Co-authored-by: mazhixin00_00<mazhixin7@huawei.com> # message auto-generated for no-merge-commit merge: !54 merge dev into dev [feature]新增RainFusionAttention算子的plugin Created-by: mazhixin00_00 Commit-by: mazhixin00_00 Merged-by: ascend-robot Description: [feature]新增RainFusionAttention算子的plugin 背景： CANN中已有稀疏算子RainFusionAttention，但是未提供python的调用方式。 方案： MindIE-SD通过注册torch算子的方式，提供算子调用方式 m.def("rainfusionattention_mindie_sd(Tensor query, Tensor key, Tensor value, Tensor select_idx, \ Tensor select_num_idx, int[] blockshape, Tensor? attn_mask=None, int[]? actual_seq_qlen=None, \ int[]? actual_seq_kvlen=None, Tensor? block_table=None, str q_input_layout='TND', str kv_input_layout='TND', \ int head_num=1, int mask_type=0, float scale=1.0, \ int inner_precise=1, int block_size=0) -> (Tensor, Tensor)"); 测试结果： 全量UT：  约束： - query,key和value目前只支持shape为(batch\*seqlen, head, headdim)的tensor - q_input_layout和kv_input_layout目前只支持TND - 该接口依赖的CANN包暂未发布，暂不支持外部调用 See merge request: Ascend/MindIE-SD!54	10 个月前
register_ops.cpp	[Feature][ops]aclnn编译工程适配 Co-authored-by: guowenna1<guowenna1@huawei.com> # message auto-generated for no-merge-commit merge: !201 merge aclnn_compil into dev [Feature][ops]aclnn编译工程适配 Created-by: guowenna1 Commit-by: guowenna1 Merged-by: ascend-robot Description: Purpose 1、算子接入方式从aclop改为aclnn，aclnn编译工程适配 2、移除rope的plugin定义，对外接口保留，但路由到torch npu实现 Test Plan A2、A3、A5编译工程测试 A2、A3全量测试用例验证 A2、A3模型整网验证 Test Report  See merge request: Ascend/MindIE-SD!201	7 个月前
rope.cpp	[dev]同步最新代码 Co-authored-by: mazhixin00_00<mazhixin7@huawei.com>	10 个月前
rope.h	[dev]同步最新代码 Co-authored-by: mazhixin00_00<mazhixin7@huawei.com>	10 个月前
sparse_block_estimate.cpp	[Feature][ops]aclnn编译工程适配 Co-authored-by: guowenna1<guowenna1@huawei.com> # message auto-generated for no-merge-commit merge: !201 merge aclnn_compil into dev [Feature][ops]aclnn编译工程适配 Created-by: guowenna1 Commit-by: guowenna1 Merged-by: ascend-robot Description: Purpose 1、算子接入方式从aclop改为aclnn，aclnn编译工程适配 2、移除rope的plugin定义，对外接口保留，但路由到torch npu实现 Test Plan A2、A3、A5编译工程测试 A2、A3全量测试用例验证 A2、A3模型整网验证 Test Report  See merge request: Ascend/MindIE-SD!201	7 个月前
sparse_block_estimate.h	[feature]新增SparseBlockEstimate算子以及plugin和UT Co-authored-by: lanwangli<lanwangli@huawei.com> # message auto-generated for no-merge-commit merge: !67 merge bsa_op1 into dev [feature]新增SparseBlockEstimate算子以及plugin和UT Created-by: mazhixin00_00 Commit-by: lanwangli Merged-by: ascend-robot Description: [feature]新增SparseBlockEstimate算子  **调用该算子plugin** import torch torch.ops.load_library("mindiesd/plugin/libPTAExtensionOPS.so") stride = 8 scale_value=query.shape[-1] ** -0.5 / stride smask, sct = torch.ops.mindie.sparse_block_estimate_mindie_sd( query=query, key=key, actual_seq_lengths=None, actual_seq_lengths_kv=None, input_layout="BNSD", stride=stride, sparse_size=128, num_heads=query.shape[1], num_key_value_heads=key.shape[1], scale_value=scale_value, threshold=1.0, causal=False, keep_sink=True, keep_recent=True, row_sparse=1.0 ) - query：支持fp16和bf16 - key：支持fp16和bf16 - actual_seq_lengths：目前只接受None - actual_seq_lengths_kv：目前只接受None - input_layout：目前只支持“BNSD”、“BSND”、“BSH” - stride：目前只支持8 - sparse_size：目前只支持128 - num_heads：query的head - num_key_value_heads：key和value的head - scale_value：headdim ** -0.5 / stride - threshold：1.0 时表示计算完整的注意力（完全不稀疏）。 - causal：为 True 时只计算 Attention Map 矩阵下三角部分，否则计算完整 Attention Map，默认值为True。 - keep_sink：为 True 时 Attention Map 矩阵的第一列必选，默认值为True。 - keep_recent：为 True 时 Attention Map 矩阵对角线上的 block 必选，默认值为True。 - row_sparse：取值为 0-1 之间的小数，用以强制保证每个 block 行的稀疏率大于 (1-row_sparse)，当 TopCDF 选择的稀疏率小于 (1-row_sparse) 时，使用 Topk进行选取，默认值为1。 **单算子测试结果：**  See merge request: Ascend/MindIE-SD!67	10 个月前