v2.11.0

New features

• Refactor op_profile to enable at runtime: #1479

Bug fixes and performance improve

• Add CI schedule trigger to runtime-build.yml: #1454
• remove execstack flag for linux: #1476
• Update calibrate_method from NoClip to Kld: #1475
• Optimize memory usage for op profile in RTOS: #1483
• optimize yolo series: #1481
• Fix side effect of softmax optimization: #1484

v2.10.0

New features

• Add k230 linux runtime python APIs and wheel package
• linux runtime support riscv gcc 14.1: #1355

Bug fixes and performance improve

• Update doc: #1224, #1240
• Add extra compile option to be compatible with k230 linux sdk: #1227
• Fix numpy non contiguous: #1232
• Upgrade conan2: #1242, #1244, #1267
• Fix sat extractor: #1243
• Fix test: #1341
• Optimize yolov10: #1260
• Optimize depth anything: #1343
• Fix rtos build isa extension: #1346
• Fix runtime dtype: #1352
• Fix k230_benchmark_test and k230_ai_demo jobs: #1359
• Fix nncase-k230 python wheel issues: #1388

nncase v2.9.0

New features

• k230 support linux runtime: #1222

Bug fixes and performance improve

• update set dotnet env: #1209
• update docs link: #1213
• fix normal reduce sum: #1214
• fix externel data: #1215

nncase v2.8.3

New features

• qemu system mode: #1201
• support layernorm channel first(C#): #1204

Bug fixes and performance improve

• Feature/add xsgetn for char array buffer: #1198
• fix reduce sum opset conditio: #1203
• Remove unavailable nuget source(nuget.cnblogs.com): #1207

nncase v2.8.2

New features

• Feature/update docs: #1174
• Feature/egraph extract constrains: #1175
• Feature/update docs: #1176

Bug fixes and performance improve

• fix fold binary: #1182
• Fix macos-latest doesn't support python 3.7 issue: #1194

nncase v2.8.1

New features

• Support multiple kmodels for accuracy_test: #1162
• Supprot npy for pytest: #1164

Bug fixes and performance improve

• Fix/dump_manager: #1163
• Fix Melgan: #1168
• Fix the missing new line issue of dynamic shape: #1170
• Fix in_ci env judge with bool type: #1171

nncase v2.8.0

New features

• Feature/gem5: #1157

Bug fixes and performance improve

• GNNE-1714:Fix/search time egraph: #1155
• Fix Not Equal : #1156
• GNNE-1714:Fix/some opt: #1158
• Fix Bucket: #1159

nncase v2.7.0

New features

• Add accuracy test: #1140
• Refactor Quantize: #1144
• Feature/k800 qemu: #1145
• GNNE-1980 Xpu Sram Codegen: #1148
• GNNE-1714:adjust for processor counter: #1150

Bug fixes and performance improve

• Fix/retinaface similarity bugs: #1141
• Fix memory leak: Avoid disposing python stream in finalizer thread: #1142
• Fix ordering of PytestCalibrationDatasetProvider: #1143
• Fix Reshape: #1147
• Fix Studio: #1151, #1152
• Add clear device mem pool util: #1153

nncase v2.6.0

New features

• Feature/cpu: #1019
• add onnx-op-support-md file: #1081
• Feature/egraph fakepass: #1111
• Feature/audio: #1118
• Feature/remove apply: #1124
• GNNE-1974: Feature/non uniform: #1128
• add CheckedShape for DistributedType: #1133
• add rvv_memcpy && opt profile: #1135
• Nncase studio: #1125, #1136, #1137

Bug fixes and performance improve

• Fix runtime: #1115
• Refactor infer report for k230_benchmark_test: #1119
• Fix/melgan: #1126
• fix LN evaluator: #1131
• Fix metadata: #1132
• GNNE-1714:Fix/regression bugs: #1116, #1117, #1121, #1123, #1127, #1130

nncase v2.5.1

New features

Bug fixes and performance improve

• Fix GetItem: #1114