diff options
| author | Tim Hall <tim.hall@arm.com> | 2021-06-08 21:25:57 +0100 |
|---|---|---|
| committer | Tim Hall <tim.hall@arm.com> | 2021-06-08 21:25:57 +0100 |
| commit | d784af7e8995a10fb403157af48371699c35bbfe (patch) | |
| tree | bf40b35b030d560049cef9411293b51e3d70ff4a /ethosu/vela/scheduler.py | |
| parent | 225e19d3640288e991475ee4c49cb3ffd83cc83b (diff) | |
| download | ethos-u-vela-d784af7e8995a10fb403157af48371699c35bbfe.tar.gz | |
MLBEDSW-4602: Fix Deepspeech scale & bias reuse issue.
- Deepspeech reuses identical weights and biases throughout
the network. Since biases are now interleaved with weights
there is a scaling issue when the ifm scales differ between
operations using the same weight and scale tensor.
- This commit uses interleaved weights/scales on their first use
but separates scales to source memory on subsequent use (if
the ifm scale is different).
Signed-off-by: Tim Hall <tim.hall@arm.com>
Change-Id: I7aae163438160a919cae04e235966e75355a6148
Diffstat (limited to 'ethosu/vela/scheduler.py')
| -rw-r--r-- | ethosu/vela/scheduler.py | 14 |
1 files changed, 11 insertions, 3 deletions
diff --git a/ethosu/vela/scheduler.py b/ethosu/vela/scheduler.py index 00a4dfc7..71007a32 100644 --- a/ethosu/vela/scheduler.py +++ b/ethosu/vela/scheduler.py @@ -94,6 +94,7 @@ class SchedulerOpInfo: self.time_index = None # Set by update_op_memory_snapshot self.ofm_depth_slices: List[int] = [0, stripe.depth] self.npu_weights_tensor = None + self.npu_scales_tensor = None self.buffered_weight_tensor = None self.cycles = None self.slack_buffering_cycles = 0 @@ -248,7 +249,10 @@ class SchedulerOperation: scheduler_op_info = SchedulerOpInfo(block_config, 0, ifm_shape, ifm2_shape, ofm_shape) if self.parent_op.weights: # Default full-depth weight encoding with no buffering - scheduler_op_info.npu_weights_tensor = weight_compressor.encode_weight_and_scale_tensor( + ( + scheduler_op_info.npu_weights_tensor, + scheduler_op_info.npu_scales_tensor, + ) = weight_compressor.encode_weight_and_scale_tensor( self.arch, self.parent_op, self.parent_op.weights, @@ -537,7 +541,7 @@ class Scheduler: ofm_full_depth_slices = [0, ref_cost.stripe.depth] # Encode weights for the full depth - full_weights = weight_compressor.encode_weight_and_scale_tensor( + full_weights, full_scales = weight_compressor.encode_weight_and_scale_tensor( self.arch, sched_op.parent_op, weight_tensor, @@ -552,9 +556,11 @@ class Scheduler: # No buffering required - take all the weights from permanent storage if sched_op.op_type == Op.FullyConnected or not needs_dma: cost.npu_weights_tensor = full_weights + cost.npu_scales_tensor = full_scales return encoded_weights = full_weights + encoded_scales = full_scales # How many NPU cycles are available under the previously executing # operator and SRAM unused for performing buffered DMA transfers @@ -609,7 +615,7 @@ class Scheduler: # Encode weights based depth slices cost.ofm_depth_slices = depth_slices - encoded_weights = weight_compressor.encode_weight_and_scale_tensor( + encoded_weights, encoded_scales = weight_compressor.encode_weight_and_scale_tensor( self.arch, sched_op.parent_op, weight_tensor, @@ -665,8 +671,10 @@ class Scheduler: # Don't slice or buffer - use the whole depth from persistent storage cost.ofm_depth_slices = ofm_full_depth_slices encoded_weights = full_weights + encoded_scales = full_scales cost.npu_weights_tensor = encoded_weights + cost.npu_scales_tensor = encoded_scales def propose_minimal_schedule(self) -> Schedule: """Proposes scheduling parameters where every operator is subdivided into the smallest stripe that satisfies the |