[QNN] Requantize - Optimize lowering for some corner cases. (#3864)
This commit is contained in:
Родитель
dee52466db
Коммит
1bc8385304
|
@ -129,6 +129,9 @@ Expr RequantizeLower(const Expr& input_tensor, const RequantizeAttrs* param,
|
|||
tensor = Subtract(tensor, input_zp);
|
||||
}
|
||||
|
||||
// If the input and output scales are same, we can skip the fixed point multiplication.
|
||||
auto scaled_int64_t = tensor;
|
||||
if (param->input_scale != param->output_scale) {
|
||||
// 3) Multiply the integer multiplier
|
||||
if (left_shift != 0) {
|
||||
tensor = Multiply(tensor, MakeConstantScalar(hp_dtype, 1 << left_shift));
|
||||
|
@ -166,11 +169,15 @@ Expr RequantizeLower(const Expr& input_tensor, const RequantizeAttrs* param,
|
|||
tensor = Add(tensor, round_scalar);
|
||||
|
||||
// 5) Simply right shift the result to get the final output.
|
||||
auto scaled_int64_t = RightShift(tensor, MakeConstantScalar(hp_dtype, total_right_shift));
|
||||
scaled_int64_t = RightShift(tensor, MakeConstantScalar(hp_dtype, total_right_shift));
|
||||
}
|
||||
|
||||
// 6) Add the output zero point.
|
||||
auto shifted_int64_t = scaled_int64_t;
|
||||
if (param->output_zero_point != 0) {
|
||||
auto output_zp = MakeConstantScalar(hp_dtype, param->output_zero_point);
|
||||
auto shifted_int64_t = Add(output_zp, scaled_int64_t);
|
||||
shifted_int64_t = Add(output_zp, scaled_int64_t);
|
||||
}
|
||||
|
||||
// 7) Clip to the out_dtype min/max.
|
||||
auto q_min = GetQmin(out_dtype);
|
||||
|
|
|
@ -64,6 +64,7 @@ def test_requantize():
|
|||
input_scale=0.5,
|
||||
output_scale=0.5,
|
||||
rounding=rounding)
|
||||
assert 'right_shift' not in mod.astext()
|
||||
verify(mod, (golden_data, golden_output))
|
||||
|
||||
def downscale_test():
|
||||
|
|
Загрузка…
Ссылка в новой задаче