diff options
| author | Eric Kunze <eric.kunze@arm.com> | 2021-10-25 16:04:20 -0700 |
|---|---|---|
| committer | Eric Kunze <eric.kunze@arm.com> | 2021-10-28 09:57:46 -0700 |
| commit | 3170439f3938d007e58998d61eed98560c3f026c (patch) | |
| tree | 6f4abac7c7c037b749c6801e95259fdd14cfd341 /chapters/data_layout.adoc | |
| parent | 82019a2621edf481e12b5de09dbfe5dc56283150 (diff) | |
| download | specification-3170439f3938d007e58998d61eed98560c3f026c.tar.gz | |
Remove zp subtraction from tensor_read pseudocode
Operators which use the zero-point functionalty for 8-bit integer
processing are updated to do the zero-point subtract in their
pseudocode.
Note that the PAD operator no longer takes a zero point argument,
and instead requires callers to account for the zero point in the
pad_const argument.
Change-Id: I3bca1cae85aa2093000c420f0433633c347a29de
Diffstat (limited to 'chapters/data_layout.adoc')
| -rw-r--r-- | chapters/data_layout.adoc | 19 |
1 files changed, 9 insertions, 10 deletions
diff --git a/chapters/data_layout.adoc b/chapters/data_layout.adoc index 205ccaf..54221f6 100644 --- a/chapters/data_layout.adoc +++ b/chapters/data_layout.adoc @@ -68,6 +68,7 @@ for_each(index1 in shape) { Pads a tensor along the borders of each dimension with a supplied value. Returns a new tensor with the padding included. +The pad_const value includes the zero point if the tensor uses a zero point. *Arguments:* @@ -77,7 +78,6 @@ Returns a new tensor with the padding included. |Input|in_t*|input1|shape1|Input tensor |Attribute|int|padding|[rank(input1),2]|Amount of padding to be done |Attribute|in_t|pad_const|-|Constant value to be used as padding -|Attribute|in_t|input1_zp|-|Input tensor zero point. Must be zero for non-int8 types. |Output|in_t*|output|shape|Output tensor of same type as the input tensor |=== @@ -85,7 +85,6 @@ Returns a new tensor with the padding included. [source,c++] ---- -ERROR_IF(in_t != int8_t && input1_zp != 0); // Zero point only allowed for int8_t // Padding sizes must be >= 0. for_each(pad_size in padding) { ERROR_IF(pad_size < 0); @@ -99,21 +98,21 @@ for_each(index in shape) { is_pad = true; } } - acc_t value = is_pad ? pad_const : tensor_read<in_t>(input1, shape1, index1, input1_zp); - tensor_write<in_t>(output, shape, index, value + input1_zp); + in_t value = is_pad ? pad_const : tensor_read<in_t>(input1, shape1, index1); + tensor_write<in_t>(output, shape, index, value); } ---- *Supported Data Types:* |=== -|Profile|Mode|in_t|acc_t +|Profile|Mode|in_t -|Any|Boolean|bool_t|bool_t -|Any|signed 8|int8_t|int16_t -|Any|signed 16|int16_t|int16_t -|Any|signed 32|int32_t|int32_t -|MI, MT|floating-point|float_t|float_t +|Any|Boolean|bool_t +|Any|signed 8|int8_t +|Any|signed 16|int16_t +|Any|signed 32|int32_t +|MI, MT|floating-point|float_t |=== ==== RESHAPE |