mirror of https://github.com/alibaba/MNN.git
1eaa7a632b
When converting a caffe model, `ReductionTransform` thinks there're always 4 dims in the input of a `Reduction` layer. However, when forwarding `OpCommonUtils::computeReduceDims` had an assumption that dim numbers wouldn't exceed `inputs[0]->dimensions()`. So, MNN would write unexpected addresses and then crash, if a Reduction op follows a Reshape / InnerProduct op, since both ops may output a tensor of 2 dimensions. _Comment_: I know `TransformInnerProduct` often converts caffe's InnerProduct into 1x1 Convolution, but Reshape can also cause a tensor has only 2 dimensions, so this still deserves a fix. |
||
|---|---|---|
| .. | ||
| AutoStorage.h | ||
| AutoTime.cpp | ||
| Backend.cpp | ||
| Backend.hpp | ||
| BackendRegister.cpp | ||
| BufferAllocator.cpp | ||
| BufferAllocator.hpp | ||
| Command.hpp | ||
| Concurrency.h | ||
| ConvolutionCommon.cpp | ||
| ConvolutionCommon.hpp | ||
| DirectedAcyclicGraph.hpp | ||
| Execution.cpp | ||
| Execution.hpp | ||
| FileLoader.cpp | ||
| FileLoader.hpp | ||
| Interpreter.cpp | ||
| MNNMemoryUtils.cpp | ||
| MNNMemoryUtils.h | ||
| Macro.h | ||
| NonCopyable.hpp | ||
| OpCommonUtils.cpp | ||
| OpCommonUtils.hpp | ||
| Pipeline.cpp | ||
| Pipeline.hpp | ||
| RuntimeFactory.cpp | ||
| RuntimeFactory.hpp | ||
| Schedule.cpp | ||
| Schedule.hpp | ||
| Session.cpp | ||
| Session.hpp | ||
| Tensor.cpp | ||
| TensorUtils.cpp | ||
| TensorUtils.hpp | ||
| WrapExecution.cpp | ||
| WrapExecution.hpp | ||