| // Copyright 2020 Google LLC |
| // |
| // This source code is licensed under the BSD-style license found in the |
| // LICENSE file in the root directory of this source tree. |
| |
| #include <assert.h> |
| #include <inttypes.h> |
| #include <stddef.h> |
| #include <stdint.h> |
| |
| #include "include/xnnpack.h" |
| #include "src/xnnpack/common.h" |
| #include "src/xnnpack/internal.h" |
| #include "src/xnnpack/log.h" |
| #include "src/xnnpack/node-type.h" |
| #include "src/xnnpack/operator-type.h" |
| #include "src/xnnpack/operator.h" |
| #include "src/xnnpack/requantization.h" |
| #include "src/xnnpack/subgraph-validation.h" |
| #include "src/xnnpack/subgraph.h" |
| #include <pthreadpool.h> |
| |
| static enum xnn_status create_deconvolution_operator( |
| const struct xnn_node* node, const struct xnn_runtime_value* values, |
| size_t num_values, struct xnn_operator_data* opdata, |
| xnn_weights_cache_t weights_cache) { |
| assert(node->num_inputs >= 2); |
| assert(node->num_inputs <= 3); |
| const bool use_bias = node->num_inputs >= 3; |
| |
| const uint32_t input_id = node->inputs[0]; |
| assert(input_id != XNN_INVALID_VALUE_ID); |
| assert(input_id < num_values); |
| const uint32_t filter_id = node->inputs[1]; |
| assert(filter_id != XNN_INVALID_VALUE_ID); |
| assert(filter_id < num_values); |
| |
| const void* bias_data = NULL; |
| uint32_t bias_id = XNN_INVALID_VALUE_ID; |
| if (use_bias) { |
| bias_id = node->inputs[2]; |
| assert(bias_id != XNN_INVALID_VALUE_ID); |
| assert(bias_id < num_values); |
| |
| bias_data = values[bias_id].fp32_data != NULL ? values[bias_id].fp32_data |
| : values[bias_id].data; |
| assert(bias_data != NULL); |
| } |
| |
| assert(node->num_outputs == 1); |
| const uint32_t output_id = node->outputs[0]; |
| assert(output_id != XNN_INVALID_VALUE_ID); |
| assert(output_id < num_values); |
| |
| const void* filter_data = values[filter_id].fp32_data != NULL |
| ? values[filter_id].fp32_data |
| : values[filter_id].data; |
| assert(filter_data != NULL); |
| |
| enum xnn_status status = xnn_status_uninitialized; |
| const enum xnn_datatype input_datatype = |
| (node->flags & XNN_FLAG_INLINE_LHS_PACKING) ? node->packed_input_datatype |
| : values[input_id].datatype; |
| const enum xnn_datatype filter_datatype = values[filter_id].datatype; |
| const enum xnn_datatype bias_datatype = bias_id != XNN_INVALID_VALUE_ID |
| ? values[bias_id].datatype |
| : xnn_datatype_invalid; |
| const enum xnn_datatype output_datatype = values[output_id].datatype; |
| switch (output_datatype) { |
| case xnn_datatype_fp16: { |
| uint32_t flags = node->flags; |
| if (filter_datatype == xnn_datatype_fp32) { |
| flags = XNN_FLAG_FP32_STATIC_WEIGHTS; |
| } |
| status = xnn_create_deconvolution2d_nhwc_f16( |
| node->params.deconvolution_2d.padding_top, |
| node->params.deconvolution_2d.padding_right, |
| node->params.deconvolution_2d.padding_bottom, |
| node->params.deconvolution_2d.padding_left, |
| node->params.deconvolution_2d.kernel_height, |
| node->params.deconvolution_2d.kernel_width, |
| node->params.deconvolution_2d.upsampling_height, |
| node->params.deconvolution_2d.upsampling_width, |
| node->params.deconvolution_2d.dilation_height, |
| node->params.deconvolution_2d.dilation_width, |
| node->params.deconvolution_2d.groups, |
| node->params.deconvolution_2d.group_input_channels, |
| node->params.deconvolution_2d.group_output_channels, |
| node->params.deconvolution_2d.group_input_channels * |
| node->params.deconvolution_2d.groups /* input_pixel_stride */, |
| node->params.deconvolution_2d.group_output_channels * |
| node->params.deconvolution_2d.groups /* output_pixel_stride */, |
| filter_data, bias_data, node->activation.output_min, |
| node->activation.output_max, flags, weights_cache, |
| &opdata->operator_objects[0]); |
| break; |
| } |
| case xnn_datatype_fp32: |
| switch (filter_datatype) { |
| case xnn_datatype_fp16: { |
| uint32_t flags = node->flags; |
| if (bias_datatype == xnn_datatype_fp32) { |
| flags |= XNN_FLAG_FP32_STATIC_BIASES; |
| } |
| status = xnn_create_deconvolution2d_nhwc_f32_f16( |
| node->params.deconvolution_2d.padding_top, |
| node->params.deconvolution_2d.padding_right, |
| node->params.deconvolution_2d.padding_bottom, |
| node->params.deconvolution_2d.padding_left, |
| node->params.deconvolution_2d.kernel_height, |
| node->params.deconvolution_2d.kernel_width, |
| node->params.deconvolution_2d.upsampling_height, |
| node->params.deconvolution_2d.upsampling_width, |
| node->params.deconvolution_2d.dilation_height, |
| node->params.deconvolution_2d.dilation_width, |
| node->params.deconvolution_2d.groups, |
| node->params.deconvolution_2d.group_input_channels, |
| node->params.deconvolution_2d.group_output_channels, |
| node->params.deconvolution_2d.group_input_channels * |
| node->params.deconvolution_2d.groups /* input_pixel_stride */, |
| node->params.deconvolution_2d.group_output_channels * |
| node->params.deconvolution_2d |
| .groups /* output_pixel_stride */, |
| filter_data, bias_data, node->activation.output_min, |
| node->activation.output_max, flags, weights_cache, |
| &opdata->operator_objects[0]); |
| break; |
| } |
| case xnn_datatype_fp32: |
| status = xnn_create_deconvolution2d_nhwc_f32( |
| node->params.deconvolution_2d.padding_top, |
| node->params.deconvolution_2d.padding_right, |
| node->params.deconvolution_2d.padding_bottom, |
| node->params.deconvolution_2d.padding_left, |
| node->params.deconvolution_2d.kernel_height, |
| node->params.deconvolution_2d.kernel_width, |
| node->params.deconvolution_2d.upsampling_height, |
| node->params.deconvolution_2d.upsampling_width, |
| node->params.deconvolution_2d.dilation_height, |
| node->params.deconvolution_2d.dilation_width, |
| node->params.deconvolution_2d.groups, |
| node->params.deconvolution_2d.group_input_channels, |
| node->params.deconvolution_2d.group_output_channels, |
| node->params.deconvolution_2d.group_input_channels * |
| node->params.deconvolution_2d.groups /* input_pixel_stride */, |
| node->params.deconvolution_2d.group_output_channels * |
| node->params.deconvolution_2d |
| .groups /* output_pixel_stride */, |
| filter_data, bias_data, node->activation.output_min, |
| node->activation.output_max, node->flags, weights_cache, |
| &opdata->operator_objects[0]); |
| break; |
| case xnn_datatype_qcint8: |
| switch (input_datatype) { |
| case xnn_datatype_qdint8: |
| status = xnn_create_deconvolution2d_nhwc_qd8_f32_qc8w( |
| node->params.deconvolution_2d.padding_top, |
| node->params.deconvolution_2d.padding_right, |
| node->params.deconvolution_2d.padding_bottom, |
| node->params.deconvolution_2d.padding_left, |
| node->params.deconvolution_2d.kernel_height, |
| node->params.deconvolution_2d.kernel_width, |
| node->params.deconvolution_2d.upsampling_height, |
| node->params.deconvolution_2d.upsampling_width, |
| node->params.deconvolution_2d.dilation_height, |
| node->params.deconvolution_2d.dilation_width, |
| node->params.deconvolution_2d.groups, |
| node->params.deconvolution_2d.group_input_channels, |
| node->params.deconvolution_2d.group_output_channels, |
| node->params.deconvolution_2d.group_input_channels * |
| node->params.deconvolution_2d |
| .groups /* input_pixel_stride */, |
| node->params.deconvolution_2d.group_output_channels * |
| node->params.deconvolution_2d |
| .groups /* output_pixel_stride */, |
| values[filter_id].quantization.channelwise_scale, filter_data, |
| bias_data, node->activation.output_min, |
| node->activation.output_max, node->flags, weights_cache, |
| &opdata->operator_objects[0]); |
| break; |
| case xnn_datatype_qduint8: |
| status = xnn_create_deconvolution2d_nhwc_qdu8_f32_qc8w( |
| node->params.deconvolution_2d.padding_top, |
| node->params.deconvolution_2d.padding_right, |
| node->params.deconvolution_2d.padding_bottom, |
| node->params.deconvolution_2d.padding_left, |
| node->params.deconvolution_2d.kernel_height, |
| node->params.deconvolution_2d.kernel_width, |
| node->params.deconvolution_2d.upsampling_height, |
| node->params.deconvolution_2d.upsampling_width, |
| node->params.deconvolution_2d.dilation_height, |
| node->params.deconvolution_2d.dilation_width, |
| node->params.deconvolution_2d.groups, |
| node->params.deconvolution_2d.group_input_channels, |
| node->params.deconvolution_2d.group_output_channels, |
| node->params.deconvolution_2d.group_input_channels * |
| node->params.deconvolution_2d |
| .groups /* input_pixel_stride */, |
| node->params.deconvolution_2d.group_output_channels * |
| node->params.deconvolution_2d |
| .groups /* output_pixel_stride */, |
| values[filter_id].quantization.channelwise_scale, filter_data, |
| bias_data, node->activation.output_min, |
| node->activation.output_max, node->flags, weights_cache, |
| &opdata->operator_objects[0]); |
| break; |
| default: |
| XNN_UNREACHABLE; |
| } |
| break; |
| default: |
| XNN_UNREACHABLE; |
| } |
| break; |
| case xnn_datatype_qint8: |
| switch (filter_datatype) { |
| case xnn_datatype_qint8: { |
| const float output_scale = values[output_id].quantization.scale; |
| const int32_t output_zero_point = |
| values[output_id].quantization.zero_point; |
| const int8_t output_min = xnn_qs8_quantize( |
| node->activation.output_min, output_scale, output_zero_point); |
| const int8_t output_max = xnn_qs8_quantize( |
| node->activation.output_max, output_scale, output_zero_point); |
| switch (input_datatype) { |
| case xnn_datatype_qint8: |
| status = xnn_create_deconvolution2d_nhwc_qs8( |
| node->params.deconvolution_2d.padding_top, |
| node->params.deconvolution_2d.padding_right, |
| node->params.deconvolution_2d.padding_bottom, |
| node->params.deconvolution_2d.padding_left, |
| node->params.deconvolution_2d.kernel_height, |
| node->params.deconvolution_2d.kernel_width, |
| node->params.deconvolution_2d.upsampling_height, |
| node->params.deconvolution_2d.upsampling_width, |
| node->params.deconvolution_2d.dilation_height, |
| node->params.deconvolution_2d.dilation_width, |
| node->params.deconvolution_2d.groups, |
| node->params.deconvolution_2d.group_input_channels, |
| node->params.deconvolution_2d.group_output_channels, |
| node->params.deconvolution_2d.group_input_channels * |
| node->params.deconvolution_2d |
| .groups /* input_pixel_stride */, |
| node->params.deconvolution_2d.group_output_channels * |
| node->params.deconvolution_2d |
| .groups /* output_pixel_stride */, |
| (int8_t)values[input_id].quantization.zero_point, |
| values[input_id].quantization.scale, |
| values[filter_id].quantization.scale, filter_data, bias_data, |
| output_zero_point, output_scale, output_min, output_max, |
| node->flags, weights_cache, &opdata->operator_objects[0]); |
| break; |
| case xnn_datatype_pqint8: |
| status = xnn_create_deconvolution2d_nhwc_pqs8_qs8_qs8( |
| node->params.deconvolution_2d.padding_top, |
| node->params.deconvolution_2d.padding_right, |
| node->params.deconvolution_2d.padding_bottom, |
| node->params.deconvolution_2d.padding_left, |
| node->params.deconvolution_2d.kernel_height, |
| node->params.deconvolution_2d.kernel_width, |
| node->params.deconvolution_2d.upsampling_height, |
| node->params.deconvolution_2d.upsampling_width, |
| node->params.deconvolution_2d.dilation_height, |
| node->params.deconvolution_2d.dilation_width, |
| node->params.deconvolution_2d.groups, |
| node->params.deconvolution_2d.group_input_channels, |
| node->params.deconvolution_2d.group_output_channels, |
| node->params.deconvolution_2d.group_input_channels * |
| node->params.deconvolution_2d |
| .groups /* input_pixel_stride */, |
| node->params.deconvolution_2d.group_output_channels * |
| node->params.deconvolution_2d |
| .groups /* output_pixel_stride */, |
| (int8_t)values[input_id].quantization.zero_point, |
| values[input_id].quantization.scale, |
| values[filter_id].quantization.scale, filter_data, bias_data, |
| output_zero_point, output_scale, output_min, output_max, |
| node->flags, weights_cache, &opdata->operator_objects[0]); |
| break; |
| default: |
| XNN_UNREACHABLE; |
| } |
| break; |
| } |
| case xnn_datatype_qcint8: { |
| const float output_scale = values[output_id].quantization.scale; |
| const int32_t output_zero_point = |
| values[output_id].quantization.zero_point; |
| const int8_t output_min = xnn_qs8_quantize( |
| node->activation.output_min, output_scale, output_zero_point); |
| const int8_t output_max = xnn_qs8_quantize( |
| node->activation.output_max, output_scale, output_zero_point); |
| switch (input_datatype) { |
| case xnn_datatype_qint8: |
| status = xnn_create_deconvolution2d_nhwc_qs8_qc8w( |
| node->params.deconvolution_2d.padding_top, |
| node->params.deconvolution_2d.padding_right, |
| node->params.deconvolution_2d.padding_bottom, |
| node->params.deconvolution_2d.padding_left, |
| node->params.deconvolution_2d.kernel_height, |
| node->params.deconvolution_2d.kernel_width, |
| node->params.deconvolution_2d.upsampling_height, |
| node->params.deconvolution_2d.upsampling_width, |
| node->params.deconvolution_2d.dilation_height, |
| node->params.deconvolution_2d.dilation_width, |
| node->params.deconvolution_2d.groups, |
| node->params.deconvolution_2d.group_input_channels, |
| node->params.deconvolution_2d.group_output_channels, |
| node->params.deconvolution_2d.group_input_channels * |
| node->params.deconvolution_2d |
| .groups /* input_pixel_stride */, |
| node->params.deconvolution_2d.group_output_channels * |
| node->params.deconvolution_2d |
| .groups /* output_pixel_stride */, |
| (int8_t)values[input_id].quantization.zero_point, |
| values[input_id].quantization.scale, |
| values[filter_id].quantization.channelwise_scale, filter_data, |
| bias_data, output_zero_point, output_scale, output_min, |
| output_max, node->flags, weights_cache, |
| &opdata->operator_objects[0]); |
| break; |
| case xnn_datatype_pqint8: |
| status = xnn_create_deconvolution2d_nhwc_pqs8_qs8_qc8w( |
| node->params.deconvolution_2d.padding_top, |
| node->params.deconvolution_2d.padding_right, |
| node->params.deconvolution_2d.padding_bottom, |
| node->params.deconvolution_2d.padding_left, |
| node->params.deconvolution_2d.kernel_height, |
| node->params.deconvolution_2d.kernel_width, |
| node->params.deconvolution_2d.upsampling_height, |
| node->params.deconvolution_2d.upsampling_width, |
| node->params.deconvolution_2d.dilation_height, |
| node->params.deconvolution_2d.dilation_width, |
| node->params.deconvolution_2d.groups, |
| node->params.deconvolution_2d.group_input_channels, |
| node->params.deconvolution_2d.group_output_channels, |
| node->params.deconvolution_2d.group_input_channels * |
| node->params.deconvolution_2d |
| .groups /* input_pixel_stride */, |
| node->params.deconvolution_2d.group_output_channels * |
| node->params.deconvolution_2d |
| .groups /* output_pixel_stride */, |
| (int8_t)values[input_id].quantization.zero_point, |
| values[input_id].quantization.scale, |
| values[filter_id].quantization.channelwise_scale, filter_data, |
| bias_data, output_zero_point, output_scale, output_min, |
| output_max, node->flags, weights_cache, |
| &opdata->operator_objects[0]); |
| break; |
| default: |
| XNN_UNREACHABLE; |
| } |
| break; |
| } |
| default: |
| XNN_UNREACHABLE; |
| } |
| break; |
| case xnn_datatype_quint8: { |
| const float output_scale = values[output_id].quantization.scale; |
| const int32_t output_zero_point = |
| values[output_id].quantization.zero_point; |
| const uint8_t output_min = xnn_qu8_quantize( |
| node->activation.output_min, output_scale, output_zero_point); |
| const uint8_t output_max = xnn_qu8_quantize( |
| node->activation.output_max, output_scale, output_zero_point); |
| status = xnn_create_deconvolution2d_nhwc_qu8( |
| node->params.deconvolution_2d.padding_top, |
| node->params.deconvolution_2d.padding_right, |
| node->params.deconvolution_2d.padding_bottom, |
| node->params.deconvolution_2d.padding_left, |
| node->params.deconvolution_2d.kernel_height, |
| node->params.deconvolution_2d.kernel_width, |
| node->params.deconvolution_2d.upsampling_height, |
| node->params.deconvolution_2d.upsampling_width, |
| node->params.deconvolution_2d.dilation_height, |
| node->params.deconvolution_2d.dilation_width, |
| node->params.deconvolution_2d.groups, |
| node->params.deconvolution_2d.group_input_channels, |
| node->params.deconvolution_2d.group_output_channels, |
| node->params.deconvolution_2d.group_input_channels * |
| node->params.deconvolution_2d.groups /* input_pixel_stride */, |
| node->params.deconvolution_2d.group_output_channels * |
| node->params.deconvolution_2d.groups /* output_pixel_stride */, |
| (uint8_t)values[input_id].quantization.zero_point, |
| values[input_id].quantization.scale, |
| (uint8_t)values[filter_id].quantization.zero_point, |
| values[filter_id].quantization.scale, filter_data, bias_data, |
| output_zero_point, output_scale, output_min, output_max, node->flags, |
| weights_cache, &opdata->operator_objects[0]); |
| break; |
| } |
| default: |
| XNN_UNREACHABLE; |
| } |
| if (status == xnn_status_success) { |
| opdata->adjustment_height = node->params.deconvolution_2d.adjustment_height; |
| opdata->adjustment_width = node->params.deconvolution_2d.adjustment_width; |
| } |
| return status; |
| } |
| |
| static enum xnn_status reshape_deconvolution_operator( |
| struct xnn_operator_data* opdata, struct xnn_runtime_value* values, |
| size_t num_values, pthreadpool_t threadpool) { |
| const uint32_t input_id = opdata->inputs[0]; |
| assert(input_id < num_values); |
| if (values[input_id].shape.num_dims != 3 && values[input_id].shape.num_dims != 4) { |
| xnn_log_error( |
| "failed to define %s operator with input ID #%" PRIu32 ": num_dims (%zu) must be 3 or 4", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), input_id, values[input_id].shape.num_dims); |
| return xnn_status_invalid_parameter; |
| } |
| const size_t batch_size = values[input_id].shape.dim[0]; |
| const size_t input_height = values[input_id].shape.dim[1]; |
| const size_t input_width = values[input_id].shape.dim[2]; |
| enum xnn_status status = xnn_status_invalid_state; |
| const size_t old_workspace_size = opdata->workspace_size; |
| size_t output_height, output_width; |
| switch (opdata->operator_objects[0]->type) { |
| case xnn_operator_type_deconvolution_nhwc_f16: |
| status = xnn_reshape_deconvolution2d_nhwc_f16( |
| opdata->operator_objects[0], batch_size, input_height, input_width, |
| opdata->adjustment_height, opdata->adjustment_width, &output_height, |
| &output_width, threadpool); |
| break; |
| case xnn_operator_type_deconvolution_nhwc_f32: |
| status = xnn_reshape_deconvolution2d_nhwc_f32( |
| opdata->operator_objects[0], batch_size, input_height, input_width, |
| opdata->adjustment_height, opdata->adjustment_width, &output_height, |
| &output_width, threadpool); |
| break; |
| case xnn_operator_type_deconvolution_nhwc_qs8: |
| status = xnn_reshape_deconvolution2d_nhwc_qs8( |
| opdata->operator_objects[0], batch_size, input_height, input_width, |
| opdata->adjustment_height, opdata->adjustment_width, &output_height, |
| &output_width, threadpool); |
| break; |
| case xnn_operator_type_deconvolution_nhwc_qs8_qc8w: |
| status = xnn_reshape_deconvolution2d_nhwc_qs8_qc8w( |
| opdata->operator_objects[0], batch_size, input_height, input_width, |
| opdata->adjustment_height, opdata->adjustment_width, &output_height, |
| &output_width, threadpool); |
| break; |
| case xnn_operator_type_deconvolution_nhwc_qu8: |
| status = xnn_reshape_deconvolution2d_nhwc_qu8( |
| opdata->operator_objects[0], batch_size, input_height, input_width, |
| opdata->adjustment_height, opdata->adjustment_width, &output_height, |
| &output_width, threadpool); |
| break; |
| case xnn_operator_type_deconvolution_nhwc_qd8_f32_qc8w: |
| status = xnn_reshape_deconvolution2d_nhwc_qd8_f32_qc8w( |
| opdata->operator_objects[0], batch_size, input_height, input_width, |
| opdata->adjustment_height, opdata->adjustment_width, &output_height, |
| &output_width, threadpool); |
| break; |
| case xnn_operator_type_deconvolution_nhwc_qdu8_f32_qc8w: |
| status = xnn_reshape_deconvolution2d_nhwc_qdu8_f32_qc8w( |
| opdata->operator_objects[0], batch_size, input_height, input_width, |
| opdata->adjustment_height, opdata->adjustment_width, &output_height, |
| &output_width, threadpool); |
| break; |
| case xnn_operator_type_deconvolution_nhwc_pqs8_qs8_qc8w: |
| status = xnn_reshape_deconvolution2d_nhwc_pqs8_qs8_qc8w( |
| opdata->operator_objects[0], batch_size, input_height, input_width, |
| opdata->adjustment_height, opdata->adjustment_width, &output_height, |
| &output_width, &opdata->workspace_size, threadpool); |
| break; |
| default: |
| XNN_UNREACHABLE; |
| } |
| if (status != xnn_status_success) { |
| return status; |
| } |
| const uint32_t output_id = opdata->outputs[0]; |
| assert(output_id < num_values); |
| struct xnn_runtime_value* output_value = values + output_id; |
| |
| const size_t output_pixel_stride = |
| opdata->operator_objects[0]->output_pixel_stride; |
| output_value->shape.dim[0] = batch_size; |
| output_value->shape.dim[1] = output_height; |
| output_value->shape.dim[2] = output_width; |
| output_value->shape.dim[3] = output_pixel_stride; |
| output_value->shape.num_dims = 4; |
| const size_t new_size = xnn_runtime_tensor_get_size(output_value); |
| if (new_size > output_value->size || |
| opdata->workspace_size > old_workspace_size) { |
| output_value->size = new_size; |
| return xnn_status_reallocation_required; |
| } |
| return xnn_status_success; |
| } |
| |
| static enum xnn_status setup_deconvolution_operator( |
| const struct xnn_operator_data* opdata, |
| const struct xnn_runtime_value* values, size_t num_values, |
| pthreadpool_t threadpool) { |
| const uint32_t input_id = opdata->inputs[0]; |
| assert(input_id != XNN_INVALID_VALUE_ID); |
| assert(input_id < num_values); |
| |
| const uint32_t output_id = opdata->outputs[0]; |
| assert(output_id != XNN_INVALID_VALUE_ID); |
| assert(output_id < num_values); |
| |
| const struct xnn_runtime_value* input_value = values + input_id; |
| const void* input_data = input_value->data; |
| assert(input_data != NULL); |
| |
| const struct xnn_runtime_value* output_value = values + output_id; |
| void* output_data = output_value->data; |
| assert(output_data != NULL); |
| |
| switch (opdata->operator_objects[0]->type) { |
| case xnn_operator_type_deconvolution_nhwc_f16: |
| return xnn_setup_deconvolution2d_nhwc_f16(opdata->operator_objects[0], |
| input_data, output_data); |
| break; |
| case xnn_operator_type_deconvolution_nhwc_f32: |
| return xnn_setup_deconvolution2d_nhwc_f32(opdata->operator_objects[0], |
| input_data, output_data); |
| break; |
| case xnn_operator_type_deconvolution_nhwc_qs8: |
| return xnn_setup_deconvolution2d_nhwc_qs8(opdata->operator_objects[0], |
| input_data, output_data); |
| break; |
| case xnn_operator_type_deconvolution_nhwc_qs8_qc8w: |
| return xnn_setup_deconvolution2d_nhwc_qs8_qc8w( |
| opdata->operator_objects[0], input_data, output_data); |
| break; |
| case xnn_operator_type_deconvolution_nhwc_qu8: |
| return xnn_setup_deconvolution2d_nhwc_qu8(opdata->operator_objects[0], |
| input_data, output_data); |
| break; |
| case xnn_operator_type_deconvolution_nhwc_qd8_f32_qc8w: { |
| const void* quantization_params = |
| input_value->quantization.dynamic_params; |
| assert(quantization_params != NULL); |
| return xnn_setup_deconvolution2d_nhwc_qd8_f32_qc8w( |
| opdata->operator_objects[0], input_data, output_data, |
| quantization_params); |
| } break; |
| case xnn_operator_type_deconvolution_nhwc_qdu8_f32_qc8w: { |
| const void* quantization_params = |
| input_value->quantization.dynamic_params; |
| assert(quantization_params != NULL); |
| return xnn_setup_deconvolution2d_nhwc_qdu8_f32_qc8w( |
| opdata->operator_objects[0], input_data, output_data, |
| quantization_params); |
| } break; |
| case xnn_operator_type_deconvolution_nhwc_pqs8_qs8_qc8w: |
| return xnn_setup_deconvolution2d_nhwc_pqs8_qs8_qc8w( |
| opdata->operator_objects[0], input_data, output_data, |
| opdata->workspace); |
| break; |
| default: |
| XNN_UNREACHABLE; |
| } |
| } |
| |
| static inline bool validate_datatypes_with_bias( |
| enum xnn_datatype input_datatype, enum xnn_datatype filter_datatype, |
| enum xnn_datatype bias_datatype, enum xnn_datatype output_datatype) { |
| switch (filter_datatype) { |
| case xnn_datatype_fp32: |
| if (input_datatype == xnn_datatype_fp32 && |
| bias_datatype == xnn_datatype_fp32 && |
| output_datatype == xnn_datatype_fp32) { |
| return true; |
| } else if (input_datatype == xnn_datatype_fp16 && |
| bias_datatype == xnn_datatype_fp32 && |
| output_datatype == xnn_datatype_fp16) { |
| // Flag: XNN_FLAG_FP32_STATIC_WEIGHTS |
| return true; |
| } |
| break; |
| case xnn_datatype_fp16: |
| if (input_datatype == xnn_datatype_fp32 && |
| bias_datatype == xnn_datatype_fp16 && |
| output_datatype == xnn_datatype_fp32) { |
| return true; |
| } else if (input_datatype == xnn_datatype_fp32 && |
| bias_datatype == xnn_datatype_fp32 && |
| output_datatype == xnn_datatype_fp32) { |
| // Flag: XNN_FLAG_FP32_STATIC_BIASES |
| return true; |
| } |
| break; |
| case xnn_datatype_qint8: |
| if (input_datatype == xnn_datatype_qint8 && |
| bias_datatype == xnn_datatype_qint32 && |
| output_datatype == xnn_datatype_qint8) { |
| return true; |
| } |
| break; |
| case xnn_datatype_quint8: |
| if (input_datatype == xnn_datatype_quint8 && |
| bias_datatype == xnn_datatype_qint32 && |
| output_datatype == xnn_datatype_quint8) { |
| return true; |
| } |
| break; |
| case xnn_datatype_qcint8: |
| if (input_datatype == xnn_datatype_qdint8 && |
| bias_datatype == xnn_datatype_fp32 && |
| output_datatype == xnn_datatype_fp32) { |
| return true; |
| } |
| if (input_datatype == xnn_datatype_qint8 && |
| bias_datatype == xnn_datatype_qcint32 && |
| output_datatype == xnn_datatype_qint8) { |
| return true; |
| } |
| break; |
| default: |
| XNN_UNREACHABLE; |
| } |
| return false; |
| } |
| |
| static inline bool validate_datatypes_without_bias( |
| enum xnn_datatype input_datatype, enum xnn_datatype filter_datatype, |
| enum xnn_datatype output_datatype) { |
| switch (filter_datatype) { |
| case xnn_datatype_fp32: |
| if (input_datatype == xnn_datatype_fp32 && |
| output_datatype == xnn_datatype_fp32) { |
| return true; |
| } else if (input_datatype == xnn_datatype_fp16 && |
| output_datatype == xnn_datatype_fp16) { |
| // Flag: XNN_FLAG_FP32_STATIC_WEIGHTS |
| return true; |
| } |
| break; |
| case xnn_datatype_fp16: |
| if (input_datatype == xnn_datatype_fp32 && |
| output_datatype == xnn_datatype_fp32) { |
| return true; |
| } |
| break; |
| case xnn_datatype_qint8: |
| if (input_datatype == xnn_datatype_qint8 && |
| output_datatype == xnn_datatype_qint8) { |
| return true; |
| } |
| break; |
| case xnn_datatype_quint8: |
| if (input_datatype == xnn_datatype_quint8 && |
| output_datatype == xnn_datatype_quint8) { |
| return true; |
| } |
| break; |
| case xnn_datatype_qcint8: |
| if (input_datatype == xnn_datatype_qdint8 && |
| output_datatype == xnn_datatype_fp32) { |
| return true; |
| } else if (input_datatype == xnn_datatype_qint8 && |
| output_datatype == xnn_datatype_qint8) { |
| return true; |
| } |
| break; |
| default: |
| XNN_UNREACHABLE; |
| } |
| return false; |
| } |
| |
| enum xnn_status xnn_define_deconvolution_2d( |
| xnn_subgraph_t subgraph, uint32_t padding_top, uint32_t padding_right, |
| uint32_t padding_bottom, uint32_t padding_left, uint32_t adjustment_height, |
| uint32_t adjustment_width, uint32_t kernel_height, uint32_t kernel_width, |
| uint32_t upsampling_height, uint32_t upsampling_width, |
| uint32_t dilation_height, uint32_t dilation_width, uint32_t groups, |
| size_t group_input_channels, size_t group_output_channels, float output_min, |
| float output_max, uint32_t input_id, uint32_t filter_id, uint32_t bias_id, |
| uint32_t output_id, uint32_t flags) { |
| enum xnn_status status; |
| if ((status = xnn_subgraph_check_xnnpack_initialized( |
| xnn_node_type_deconvolution_2d)) != xnn_status_success) { |
| return status; |
| } |
| |
| if (kernel_width == 0 || kernel_height == 0) { |
| xnn_log_error("failed to define %s operator with %" PRIu32 "x%" PRIu32 |
| " kernel: kernel dimensions must be non-zero", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| kernel_width, kernel_height); |
| return xnn_status_invalid_parameter; |
| } |
| |
| if (upsampling_width == 0 || upsampling_height == 0) { |
| xnn_log_error("failed to define %s operator with %" PRIu32 "x%" PRIu32 |
| " upsampling: upsampling dimensions must be non-zero", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| upsampling_width, upsampling_height); |
| return xnn_status_invalid_parameter; |
| } |
| |
| if (dilation_width == 0 || dilation_height == 0) { |
| xnn_log_error("failed to define %s operator with %" PRIu32 "x%" PRIu32 |
| " dilation: dilation dimensions must be non-zero", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| dilation_width, dilation_height); |
| return xnn_status_invalid_parameter; |
| } |
| |
| if (groups == 0) { |
| xnn_log_error("failed to define %s operator with %" PRIu32 |
| " groups: number of groups must be non-zero", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| groups); |
| return xnn_status_invalid_parameter; |
| } |
| |
| if (group_input_channels == 0) { |
| xnn_log_error( |
| "failed to define %s operator with %zu input channels per group: " |
| "number of channels must be non-zero", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| group_input_channels); |
| return xnn_status_invalid_parameter; |
| } |
| |
| if (group_output_channels == 0) { |
| xnn_log_error( |
| "failed to define %s operator with %zu output channels per group: " |
| "number of channels must be non-zero", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| group_output_channels); |
| return xnn_status_invalid_parameter; |
| } |
| |
| status = xnn_subgraph_check_output_min_max(xnn_node_type_deconvolution_2d, |
| output_min, output_max); |
| if (status != xnn_status_success) { |
| return status; |
| } |
| |
| if ((status = xnn_subgraph_check_input_node_id( |
| xnn_node_type_deconvolution_2d, input_id, subgraph->num_values)) != |
| xnn_status_success) { |
| return status; |
| } |
| |
| const struct xnn_value* input_value = &subgraph->values[input_id]; |
| status = xnn_subgraph_check_input_type_dense(xnn_node_type_deconvolution_2d, |
| input_id, input_value); |
| if (status != xnn_status_success) { |
| return status; |
| } |
| |
| switch (input_value->datatype) { |
| case xnn_datatype_fp16: |
| case xnn_datatype_fp32: |
| case xnn_datatype_qint8: |
| case xnn_datatype_quint8: |
| break; |
| case xnn_datatype_qdint8: |
| if (input_value->quantization.num_nonbatch_dims >= |
| input_value->shape.num_dims) { |
| xnn_log_error("failed to define %s operator with input ID #%" PRIu32 |
| ": num_nonbatch_dims (%zu) must be < num_dims (%zu)", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| input_id, input_value->quantization.num_nonbatch_dims, |
| input_value->shape.num_dims); |
| return xnn_status_invalid_parameter; |
| } |
| break; |
| default: |
| xnn_log_error("failed to define %s operator with input ID #%" PRIu32 |
| ": unsupported Value datatype %s (%d)", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| input_id, xnn_datatype_to_string(input_value->datatype), |
| input_value->datatype); |
| return xnn_status_invalid_parameter; |
| } |
| |
| if (filter_id >= subgraph->num_values) { |
| xnn_log_error("failed to define %s operator with filter ID #%" PRIu32 |
| ": invalid Value ID", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| filter_id); |
| return xnn_status_invalid_parameter; |
| } |
| |
| const struct xnn_value* filter_value = &subgraph->values[filter_id]; |
| if (filter_value->type != xnn_value_type_dense_tensor) { |
| xnn_log_error("failed to define %s operator with filter ID #%" PRIu32 |
| ": unsupported Value type %d (expected dense tensor)", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| filter_id, filter_value->type); |
| return xnn_status_invalid_parameter; |
| } |
| |
| if (filter_value->data == NULL) { |
| xnn_log_error("failed to define %s operator with filter ID #%" PRIu32 |
| ": non-static Value", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| filter_id); |
| return xnn_status_invalid_parameter; |
| } |
| |
| switch (filter_value->datatype) { |
| case xnn_datatype_fp16: |
| case xnn_datatype_fp32: |
| break; |
| case xnn_datatype_qcint8: |
| case xnn_datatype_qint8: |
| if (filter_value->quantization.zero_point != 0) { |
| xnn_log_error("failed to define %s operator with filter ID #%" PRIu32 |
| ": unsupported quantization zero point %" PRId32 |
| " for datatype %s", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| filter_id, filter_value->quantization.zero_point, |
| xnn_datatype_to_string(filter_value->datatype)); |
| } |
| break; |
| case xnn_datatype_quint8: |
| break; |
| default: |
| xnn_log_error("failed to define %s operator with filter ID #%" PRIu32 |
| ": unsupported Value datatype %s (%d)", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| filter_id, xnn_datatype_to_string(filter_value->datatype), |
| filter_value->datatype); |
| return xnn_status_invalid_parameter; |
| } |
| |
| const struct xnn_value* bias_value = NULL; |
| |
| if (bias_id != XNN_INVALID_VALUE_ID) { |
| if (bias_id >= subgraph->num_values) { |
| xnn_log_error("failed to define %s operator with bias ID #%" PRIu32 |
| ": invalid Value ID", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| bias_id); |
| return xnn_status_invalid_parameter; |
| } |
| |
| bias_value = &subgraph->values[bias_id]; |
| if (bias_value->type != xnn_value_type_dense_tensor) { |
| xnn_log_error("failed to define %s operator with bias ID #%" PRIu32 |
| ": unsupported Value type %d (expected dense tensor)", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| bias_id, bias_value->type); |
| return xnn_status_invalid_parameter; |
| } |
| |
| if (bias_value->data == NULL) { |
| xnn_log_error("failed to define %s operator with bias ID #%" PRIu32 |
| ": non-static Value", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| bias_id); |
| return xnn_status_invalid_parameter; |
| } |
| |
| switch (bias_value->datatype) { |
| case xnn_datatype_fp16: |
| case xnn_datatype_fp32: |
| case xnn_datatype_qint32: |
| case xnn_datatype_qcint32: |
| break; |
| default: |
| xnn_log_error("failed to define %s operator with bias ID #%" PRIu32 |
| ": unsupported Value datatype %s (%d)", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| bias_id, xnn_datatype_to_string(bias_value->datatype), |
| bias_value->datatype); |
| return xnn_status_invalid_parameter; |
| } |
| } |
| |
| status = xnn_subgraph_check_output_node_id(xnn_node_type_deconvolution_2d, |
| output_id, subgraph->num_values); |
| if (status != xnn_status_success) { |
| return status; |
| } |
| |
| const struct xnn_value* output_value = &subgraph->values[output_id]; |
| status = xnn_subgraph_check_output_type_dense(xnn_node_type_deconvolution_2d, |
| output_id, output_value); |
| if (status != xnn_status_success) { |
| return status; |
| } |
| |
| switch (output_value->datatype) { |
| case xnn_datatype_fp16: |
| case xnn_datatype_fp32: |
| case xnn_datatype_qint8: |
| case xnn_datatype_quint8: |
| break; |
| default: |
| xnn_log_error("failed to define %s operator with output ID #%" PRIu32 |
| ": unsupported Value datatype %s (%d)", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| output_id, xnn_datatype_to_string(output_value->datatype), |
| output_value->datatype); |
| return xnn_status_invalid_parameter; |
| } |
| |
| if (bias_value != NULL) { |
| if (!validate_datatypes_with_bias( |
| input_value->datatype, filter_value->datatype, bias_value->datatype, |
| output_value->datatype)) { |
| xnn_log_error("failed to define %s operator with input ID #%" PRIu32 |
| ", filter ID #%" PRIu32 ", bias ID #%" PRIu32 |
| ", and output ID #%" PRIu32 |
| ": mismatching datatypes across input (%s), filter (%s), " |
| "bias (%s), and output (%s)", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| input_id, filter_id, bias_id, output_id, |
| xnn_datatype_to_string(input_value->datatype), |
| xnn_datatype_to_string(filter_value->datatype), |
| xnn_datatype_to_string(bias_value->datatype), |
| xnn_datatype_to_string(output_value->datatype)); |
| return xnn_status_invalid_parameter; |
| } |
| } else { |
| if (!validate_datatypes_without_bias(input_value->datatype, |
| filter_value->datatype, |
| output_value->datatype)) { |
| xnn_log_error("failed to define %s operator with input ID #%" PRIu32 |
| ", filter ID #%" PRIu32 ", and output ID #%" PRIu32 |
| ": mismatching datatypes across input (%s), filter (%s), " |
| "and output (%s)", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| input_id, filter_id, output_id, |
| xnn_datatype_to_string(input_value->datatype), |
| xnn_datatype_to_string(filter_value->datatype), |
| xnn_datatype_to_string(output_value->datatype)); |
| return xnn_status_invalid_parameter; |
| } |
| } |
| |
| if (filter_value->datatype == xnn_datatype_qcint8) { |
| if (filter_value->quantization.channel_dimension != 0) { |
| xnn_log_error("failed to define %s operator with filter ID #%" PRIu32 |
| ": invalid channel dimension %zu", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| filter_id, filter_value->quantization.channel_dimension); |
| return xnn_status_invalid_parameter; |
| } |
| |
| if (bias_value != NULL) { |
| assert(bias_value->datatype == xnn_datatype_qcint32 || |
| bias_value->datatype == xnn_datatype_fp32); |
| if (bias_value->datatype == xnn_datatype_qcint32 && |
| bias_value->quantization.channel_dimension != 0) { |
| xnn_log_error("failed to define %s operator with bias ID #%" PRIu32 |
| ": invalid channel dimension %zu", |
| xnn_node_type_to_string(xnn_node_type_deconvolution_2d), |
| bias_id, bias_value->quantization.channel_dimension); |
| return xnn_status_invalid_parameter; |
| } |
| } |
| } |
| |
| struct xnn_node* node = xnn_subgraph_new_node(subgraph); |
| if (node == NULL) { |
| return xnn_status_out_of_memory; |
| } |
| |
| node->type = xnn_node_type_deconvolution_2d; |
| node->params.deconvolution_2d.padding_top = padding_top; |
| node->params.deconvolution_2d.padding_right = padding_right; |
| node->params.deconvolution_2d.padding_bottom = padding_bottom; |
| node->params.deconvolution_2d.padding_left = padding_left; |
| node->params.deconvolution_2d.kernel_height = kernel_height; |
| node->params.deconvolution_2d.kernel_width = kernel_width; |
| node->params.deconvolution_2d.upsampling_height = upsampling_height; |
| node->params.deconvolution_2d.upsampling_width = upsampling_width; |
| node->params.deconvolution_2d.dilation_height = dilation_height; |
| node->params.deconvolution_2d.dilation_width = dilation_width; |
| node->params.deconvolution_2d.adjustment_height = adjustment_height; |
| node->params.deconvolution_2d.adjustment_width = adjustment_width; |
| node->params.deconvolution_2d.groups = groups; |
| node->params.deconvolution_2d.group_input_channels = group_input_channels; |
| node->params.deconvolution_2d.group_output_channels = group_output_channels; |
| node->activation.output_min = output_min; |
| node->activation.output_max = output_max; |
| node->num_inputs = 2 + (size_t)(bias_value != NULL); |
| node->inputs[0] = input_id; |
| node->inputs[1] = filter_id; |
| node->inputs[2] = bias_id; |
| node->num_outputs = 1; |
| node->outputs[0] = output_id; |
| node->flags = flags; |
| |
| node->create = create_deconvolution_operator; |
| node->reshape = reshape_deconvolution_operator; |
| node->setup = setup_deconvolution_operator; |
| |
| return xnn_status_success; |
| }; |
