shoubhik commented on a change in pull request #5153: Adding support for QNN subtract op URL: https://github.com/apache/incubator-tvm/pull/5153#discussion_r399443268
########## File path: src/relay/qnn/op/subtract.cc ########## @@ -0,0 +1,103 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, + * software distributed under the License is distributed on an + * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + * KIND, either express or implied. See the License for the + * specific language governing permissions and limitations + * under the License. + */ + +/*! + * \file src/relay/qnn/op/subtract.cc + * \brief QNN add operator. + */ +#include <tvm/relay/analysis.h> +#include <tvm/relay/op_attr_types.h> +#include "op_common.h" + +namespace tvm { +namespace relay { +namespace qnn { + +/* + * \brief Canonicalizes the QNN subtract op. + * \param attrs The empty attribute. + * \param new_args The new mutated args to the call node. + * \param arg_types The types of input and output. + * \return The sequence of Relay ops for add op. + */ +Expr QnnSubtractCanonicalize(const Attrs &attrs, + const Array<Expr> &new_args, + const Array<tvm::relay::Type> &arg_types) { + // Get the args. + QnnBinaryOpArguments args(new_args); + + // Get the input dtype and shape. + QnnBinaryOpTensorType input_type(arg_types, 0); + + // TODO(shoubhik) - The lowering can be further optimized. Instead of inserting requantize in + // the start, we can insert requantize at the end if both input tensors have same qnn params. In + // that case, we can first subtract the tensors, add the zero point, and requantize at the end. + // This can be done in future. + + // Since the input qnn params can be different than output qnn params, we first requantize the + // input tensors to the output qnn params. Then we call relay.subtract on the requantized inputs. + // This subtraction results in extra subtraction of the output zero point. We further add + // the zero point. The whole process can be represented using following equations + // + // scale_c * (Q_c - zp_c) = scale_a * (Q_a - zp_a) - scale_b * (Q_b - zp_b) + // + // After requantizing Q_a and Q_b, equation becomes, + // scale_c * (Q_c - zp_c) = scale_c * (Q_a' - zp_c) - scale_c * (Q_b' - zp_c) + // scale_c * (Q_c - zp_c) = scale_c * (Q_a' - Q_b') + // + // Comparing the LHS and RHS, it results in + // Q_c = Q_a' - Q_b' + zp_c + // The subtract op is done in int32 precision. + + // Requantize LHS if necessary. Computes Q_a' + auto requantized_lhs = RequantizeOrUpcast(args.lhs, args.lhs_scale, + args.lhs_zero_point, + args.output_scale, + args.output_zero_point, + input_type.shape); + // Requantize RHS if necessary. Computes Q_b' + auto requantized_rhs = RequantizeOrUpcast(args.rhs, args.rhs_scale, + args.rhs_zero_point, + args.output_scale, + args.output_zero_point, + input_type.shape); + + // Computes Q_a' - Q_b' + auto output = Subtract(requantized_lhs, requantized_rhs); + + // Add zero point. Computes (Q_a' - Q_b') + zp_c + auto zero_scalar = MakeConstantScalar(DataType::Int(32), 0); + if (!IsEqualScalar(args.output_zero_point, zero_scalar)) { + output = Add(output, args.output_zero_point); + } + + // Go back to lower precision. + return ConvertDtype(output, input_type.dtype); +} + +// QNN Addition operator. +QNN_REGISTER_BINARY_OP("subtract") +.describe("Elementwise subtract with with broadcasting for quantized tensors.") +.set_support_level(11) +.set_attr<FTVMLegalize>("FTVMQnnCanonicalize", QnnSubtractCanonicalize) +.set_attr<FInferCorrectLayout>("FInferCorrectLayout", QnnBinaryBroadcastLayout); Review comment: I beleive this should be here. otherwise there woluld be a confusion as to where the op is registered. I have seen this convention being followed in all other cc files. I think we should stick with the convention. plus, there is a danger of moving this to .h file. just is case someone screws up later and removes or forgets to link the header file these ops will never be registerd. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
