diff options
| author | Fredrik Svedberg <fredrik.svedberg@arm.com> | 2023-03-09 13:22:40 +0100 |
|---|---|---|
| committer | Fredrik Svedberg <fredrik.svedberg@arm.com> | 2023-03-13 15:44:32 +0000 |
| commit | bb9885190f5f7ea959f171b38ee1dd44d3e1e75e (patch) | |
| tree | ad87c79350f14e56760903f6da2dc1ca107928b3 | |
| parent | 6e281afe19ea0cd9dba2cecfb73050c18f29d242 (diff) | |
| download | ethos-u-vela-bb9885190f5f7ea959f171b38ee1dd44d3e1e75e.tar.gz | |
MLBEDSW-7427 Fix scale calculations for FullyConnected
Fixed scale calculations for FullyConnected to match the reference.
Also removed unused low_precision_scaling.
Change-Id: I4b766febff4a0010acd3de708bb49be458d22bf3
Signed-off-by: Fredrik Svedberg <fredrik.svedberg@arm.com>
| -rw-r--r-- | ethosu/vela/operation.py | 5 | ||||
| -rw-r--r-- | ethosu/vela/weight_compressor.py | 14 |
2 files changed, 4 insertions, 15 deletions
diff --git a/ethosu/vela/operation.py b/ethosu/vela/operation.py index f85cb4bb..19b00b31 100644 --- a/ethosu/vela/operation.py +++ b/ethosu/vela/operation.py @@ -487,7 +487,6 @@ class Operation: "read_shapes", "rounding_mode", "explicit_scaling", - "low_precision_scaling", "write_offset", "write_shape", "ifm_resampling_mode", @@ -525,9 +524,6 @@ class Operation: self.rounding_mode: Optional[NpuRoundingMode] = None # Rescale op in TOSA supplies explicit multiplier and shift values self.explicit_scaling: Optional[ExplicitScaling] = None - # The Mean operator (implemented as a depthwise convolution) requires scaling - # to be calculated differently in one case. In that case, this is set to True. - self.low_precision_scaling = False # Write offset, for operations that only produce a part of the OFM self.write_offset: Optional[Shape4D] = None # The amount of OFM that is produced by the operation (only if write_offset is not None). @@ -567,7 +563,6 @@ class Operation: res.write_shape = Shape4D(*self.write_shape) if self.write_shape else None res.rounding_mode = self.rounding_mode res.explicit_scaling = self.explicit_scaling - res.low_precision_scaling = self.low_precision_scaling res.ifm_resampling_mode = self.ifm_resampling_mode res.tile_base_offsets_ifm = [_ifm.copy() for _ifm in self.tile_base_offsets_ifm] res.tile_base_offsets_ofm = self.tile_base_offsets_ofm.copy() diff --git a/ethosu/vela/weight_compressor.py b/ethosu/vela/weight_compressor.py index e56cc5e5..ab22e94f 100644 --- a/ethosu/vela/weight_compressor.py +++ b/ethosu/vela/weight_compressor.py @@ -266,17 +266,11 @@ def _prepare_scale_and_bias(arch, tens, rescale_for_faf, explicit_scaling): # Convert scales to np.double (from np.float32) to conform to TensorFlow Lite which # uses double during scaling calculations - # TensorFlow Lite casts the scales slightly differently for uint8 and int8 + # TensorFlow Lite casts the scales slightly differently for uint8 and int8 as well as + # for FullyConnected operators if not rescale_for_faf: - if ifm_dtype == DataType.uint8: - # for some cases of the Mean operator, the scale must be calculated differently to match reference - if first_consumer_op.low_precision_scaling: - scales = [ - np.double(np.single(ifm_scale) / (np.single(weight_scale) * np.single(ofm_scale))) - for weight_scale in weight_scales - ] - else: - scales = [np.double(ifm_scale * weight_scale) / np.double(ofm_scale) for weight_scale in weight_scales] + if ifm_dtype == DataType.uint8 or first_consumer_op.type == Op.FullyConnected: + scales = [np.double(ifm_scale * weight_scale) / np.double(ofm_scale) for weight_scale in weight_scales] elif ifm_dtype == DataType.int8 or ifm_dtype == DataType.int16: scales = [ (np.double(ifm_scale) * np.double(weight_scale)) / np.double(ofm_scale) |