2.39.0 |
|
2.38.0 |
Support GatherNd Operation.
Support INT_32 datatype for Argmax and Argmin operation
Support UINT_32 datatype for Argmax and Argmin operation
Support CumulativeSum Operation.
|
2.34.0 |
|
2.32.0 |
|
2.28.0 |
|
2.26.0 |
|
2.25.0 |
|
2.23.0 |
|
2.22.0 |
Support QNN_DATATYPE_UFIXED_POINT_4 in ElementwiseSelect op.
Support QNN_DATATYPE_UFIXED_POINT_4 in BinaryElementwise ops.
Support Concat op with input rank 5 and axis = 1 if batch = 1
Support QNN_DATATYPE_UFIXED_POINT_4 in Gather op.
|
2.21.0 |
Support QNN_DATATYPE_BOOL_8 inputs/outputs to Reshape op.
Support QNN_DATATYPE_BOOL_8 outputs to Cast op.
Support QNN_DATATYPE_INT_32 in Concat op.
Support QNN_DATATYPE_INT_32 in BinaryElementwise op.
Support QNN_DATATYPE_UFIXED_POINT_4 in Dequantize op.
Support ScatterND Operation.
|
2.20.0 |
|
2.18.0 |
|
2.16.0 |
Support 5D inputs in StridedSlice operation.
Support Concat op for input rank = 5 and axis > 1.
Support ElementWiseNeuron Operation
|
2.14.0 |
Support BOOL_8 inputs to Cast operation.
Support CRD Mode in DepthToSpace Operation.
Support ElementWiseUnary Operation
Support ElementWiseBinary Operation
|
2.13.0 |
Support LayerNorm for input rank < 4 and batch > 1 for rank 4 input.
Support Batch Broadcasting in MatMul Operation.
Support QNN_DATATYPE_INT_64 input datatype in Cast Operation.
Support reduction across batch
Support Broadcasting for ElementwiseSelect Operation.
|
2.12.0 |
|
2.10.0 |
|
2.9.0 |
|
2.3.0 |
Support INT_8 inputs/outputs in Cast Operation.
Support UINT_32 inputs/outputs in StridedSlice Operation.
Support INT_32 inputs/outputs in Transpose Operation.
Support UINT_8 inputs/outputs in Reshape Operation.
Update Tile operation validation to disallow tiling across batch dimension.
|
2.0.0 |
|