CPUExecutionProvider
    unchanged

CUDAExecutionProvider
    unchanged

CoreMLExecutionProvider
    +2 MatMul
    +2 Mul
    -1 Attention is_causal=0 qk_matmul_output_mode=0 softcap=0.0
    -1 ReduceL2 keepdims=1 noop_with_empty_axes=0
    +1 ReduceSum keepdims=1
    +1 Softmax axis=-1
    +1 Sqrt
    +1 Transpose perm=(0, 1, 3, 2)
    -1 Transpose perm=(0, 3, 1, 2)
    239 nodes, weights 6f33ed21a226
    DecomposeAttention x5 b00b8abe
    DecomposeReduceL2 x3 31703472
    FuseBatchNormIntoConv, RemoveOptionalBiasFromConv x85 7e00f748
    _ConstantifyReshapeTarget x9 8c4eb3c1
    _ScalarGatherToSlice x1 72ff007c
    unstamped x136 157e484d

MIGraphXExecutionProvider
    +2 MatMul
    -1 Attention is_causal=0 qk_matmul_output_mode=0 softcap=0.0
    +1 Mul
    +1 Softmax axis=-1
    +1 Transpose perm=(0, 1, 3, 2)
    238 nodes, weights 6f33ed21a226
    DecomposeAttention x5 b00b8abe
    FuseBatchNormIntoConv, RemoveOptionalBiasFromConv x85 1ad75ae5
    _ConstantifyReshapeTarget x9 8c4eb3c1
    _ScalarGatherToSlice x1 72ff007c
    unstamped x138 743040bd

NvTensorRTRTXExecutionProvider
    +2 MatMul
    -1 Attention is_causal=0 qk_matmul_output_mode=0 softcap=0.0
    +1 Mul
    +1 Softmax axis=-1
    +1 Transpose perm=(0, 1, 3, 2)
    238 nodes, weights 6f33ed21a226
    DecomposeAttention x5 b00b8abe
    FuseBatchNormIntoConv, RemoveOptionalBiasFromConv x85 1ad75ae5
    _ConstantifyReshapeTarget x9 8c4eb3c1
    _ScalarGatherToSlice x1 72ff007c
    unstamped x138 743040bd

OpenVINOExecutionProvider
    +2 MatMul
    -1 Attention is_causal=0 qk_matmul_output_mode=0 softcap=0.0
    +1 Mul
    +1 Softmax axis=-1
    +1 Transpose perm=(0, 1, 3, 2)
    238 nodes, weights 6f33ed21a226
    DecomposeAttention x5 b00b8abe
    FuseBatchNormIntoConv, RemoveOptionalBiasFromConv x85 1ad75ae5
    _ConstantifyReshapeTarget x9 8c4eb3c1
    _ScalarGatherToSlice x1 72ff007c
    unstamped x138 743040bd

RKNPU static
    +33 Slice
    +24 Add
    +14 MatMul
    +12 Conv auto_pad=NOTSET dilations=(1, 1) group=1 pads=(1, 1, 1, 1) strides=(1, 1)
    -1 Attention is_causal=0 qk_matmul_output_mode=0 softcap=0.0
    -1 Cast to=1
    +1 Mul
    +1 Softmax axis=-1
    +1 Transpose perm=(0, 1, 3, 2)
    318 nodes, weights b07e1d97dd4f
    contract {"dims":[{}]}
    opset 19
    DecomposeAttention x5 b00b8abe
    FuseBatchNormIntoConv, RemoveOptionalBiasFromConv x79 f89f4aa0
    SplitLargeConvReduction, FuseBatchNormIntoConv, RemoveOptionalBiasFromConv x48 db035233
    SplitLargeReduction x42 fe862331
    _ConstantifyReshapeTarget x9 8c4eb3c1
    _ScalarGatherToSlice x1 72ff007c
    unstamped x134 38b5d3b6

TensorrtExecutionProvider
    +2 MatMul
    -1 Attention is_causal=0 qk_matmul_output_mode=0 softcap=0.0
    +1 Mul
    +1 Softmax axis=-1
    +1 Transpose perm=(0, 1, 3, 2)
    238 nodes, weights 6f33ed21a226
    DecomposeAttention x5 b00b8abe
    FuseBatchNormIntoConv, RemoveOptionalBiasFromConv x85 1ad75ae5
    _ConstantifyReshapeTarget x9 8c4eb3c1
    _ScalarGatherToSlice x1 72ff007c
    unstamped x138 743040bd
