| COMPMID-1676: Change CLROIAlign interface to accept ROIs as tensors |
|
More...
|
about 8 years ago
|
| COMPMID-1681: (Nightly) NEWidthConcatenateLayer fails |
|
More...
|
about 8 years ago
|
| COMPMID-1626: Fixed VGG 16/19 bad_alloc failure. |
|
More...
|
about 8 years ago
|
| COMPMID-1451 Properly remove dimensions in CLReduceMean |
|
More...
|
about 8 years ago
|
| COMPMID-1673: Collapse window in CLArithmeticAddition when one operand is a vector |
|
More...
|
about 8 years ago
|
| COMPMID-1638 - Add BBoxTransform to the graph API |
|
More...
|
about 8 years ago
|
| COMPMID-1029: Collapse CLWinogradInputTransform/CLWinogradOutputTransform |
|
More...
|
about 8 years ago
|
| COMPMID-1451: Fuse activation in DepthwiseConvolution. |
|
More...
|
about 8 years ago
|
| COMPMID-1451: Perform CLOutputStage using floats. |
|
More...
|
about 8 years ago
|
| COMPMID-1667: Add 4D tensors support to CLWidthConcatenateLayerKernel |
|
More...
|
about 8 years ago
|
| COMPMID-1637 - Add ROIAlign operator to the graph API |
|
More...
|
about 8 years ago
|
| COMPMID-1661 Nightly - CLReductionOperation FP16 failures |
|
More...
|
about 8 years ago
|
| COMPMID-1451 - Enabled NHWC in AlexNet, InceptionV3 and InceptionV4 |
|
More...
|
about 8 years ago
|
| COMPMID-1413 - Improve the performance of GEMMLowp with 8 bit dot product on OpenCL COMPMID-1424 - Add dot product support for CLDepthwise QASYMM8 3x3 NHWC non-unit stride |
|
More...
|
about 8 years ago
|
| COMPMID-1659: (Nightly) CLGEMMConvolutionLayer QASYMM8 TensorShape error |
|
More...
|
about 8 years ago
|
| COMPMID-1451: Fix allocation of weights in Deconvolution |
|
More...
|
about 8 years ago
|
| COMPMID-1451:Fuse RELU,LU_BOUNDED_RELU with requantization in NEGEMMConvolutionLayer. |
|
More...
|
about 8 years ago
|
| COMPMID-1657 - Fix CPU info in the json file output |
|
More...
|
about 8 years ago
|
| COMPMID-1580 Implement ReduceMean in NEON |
|
More...
|
about 8 years ago
|
| COMPMID-1451: Reverting changes for CLGEMM and CLGEMMLowp previuosly done (384496) Mirroring CLGEMM behaviour to CLGEMMLowp |
|
More...
|
about 8 years ago
|
| COMPMID-1451: Fix NormalizationLayer accross width normalization. |
|
More...
|
about 8 years ago
|
| COMPMID-1632 Add CLL2NormalizationLayer for NHWC and FP32 |
|
More...
|
about 8 years ago
|
| COMPMID-1605: API alignment for the MemoryManager with ARMNN |
|
More...
|
about 8 years ago
|
| COMPMID-1631: Fixed test framework memory corruption. |
|
More...
|
about 8 years ago
|
| COMPMID-1622 : (Nightly) Seg fault on android |
|
More...
|
about 8 years ago
|
| COMPMID-1451: Fix compilation issues under gcc 8 |
|
More...
|
about 8 years ago
|
| COMPMID-1451: Enable dot kernels in NEGEMMAssembly functions |
|
More...
|
about 8 years ago
|
| COMPMID-1630: Fixed broken DOTPROD detection. |
|
More...
|
about 8 years ago
|
| COMPMID-1628: NEWinograd mismatches in the nightly. |
|
More...
|
about 8 years ago
|
| COMPMID-1622 : (Nightly) Seg fault on android |
|
More...
|
about 8 years ago
|