format

PaddlePaddle · Mar 3, 2023 · 512600b · 512600b
1 parent 2cd3a88
commit 512600b
Show file tree

Hide file tree

Showing 3 changed files with 2 additions and 70 deletions.
diff --git a/lite/kernels/xpu/__xpu__spatial_transformer_compute.cc b/lite/kernels/xpu/__xpu__spatial_transformer_compute.cc
@@ -96,8 +96,6 @@ template <typename InType, PrecisionType PType>
 void XPUSpatialTransformerCompute<InType, PType>::PrepareForRun() {
   auto& ctx = this->ctx_->template As<XPUContext>();
   auto& param = this->template Param<param_t>();
-// prepare fc bias
-#ifdef USE_XFT
   xft_attn_fc_bias.emplace_back(
       const_cast<float*>(param.fc_bias[0]->template data<float>()),
       xft::xftVec<float>::dim_t{param.fc_bias[0]->dims()[0]});
@@ -110,69 +108,44 @@ void XPUSpatialTransformerCompute<InType, PType>::PrepareForRun() {
   xft_geglu_fc_bias.emplace_back(
       const_cast<float*>(param.fc_bias[3]->template data<float>()),
       xft::xftVec<float>::dim_t{param.fc_bias[3]->dims()[0]});
-#else
-  for (auto* fc_bias : param.fc_bias) {
-    arg_fc_bias_.push_back(fc_bias->template data<float>());
-  }
-#endif
   // prepare scale
   for (auto* ln_scale : param.ln_scale) {
-#ifdef USE_XFT
     xft_ln_weights.emplace_back(
         const_cast<float*>(ln_scale->template data<float>()),
         xft::xftVec<float>::dim_t{ln_scale->dims()[0]});
-#else
-    arg_ln_scale_.push_back(ln_scale->template data<float>());
-#endif
   }
   // prepare ln_bias
   for (auto* ln_bias : param.ln_bias) {
-#ifdef USE_XFT
     xft_ln_bias.emplace_back(
         const_cast<float*>(ln_bias->template data<float>()),
         xft::xftVec<float>::dim_t{ln_bias->dims()[0]});
-#else
-    arg_ln_bias_.push_back(ln_bias->template data<float>());
-#endif
   }
 
   // prepare gn_scale
   for (auto* gn_scale : param.gn_scale) {
-#ifdef USE_XFT
     xft_gn_weights.emplace_back(
         const_cast<float*>(gn_scale->template data<float>()),
         xft::xftVec<float>::dim_t{gn_scale->dims()[0]});
-#else
-    arg_gn_scale_.push_back(gn_scale->template data<float>());
-#endif
   }
   // prepare gn_bias
   for (auto* gn_bias : param.gn_bias) {
-#ifdef USE_XFT
     xft_gn_bias.emplace_back(
         const_cast<float*>(gn_bias->template data<float>()),
         xft::xftVec<float>::dim_t{gn_bias->dims()[0]});
-#else
-    arg_gn_bias_.push_back(gn_bias->template data<float>());
-#endif
   }
   // prepare conv bias
   for (auto* conv_bias : param.conv_bias) {
-#ifdef USE_XFT
     xft_conv_bias.emplace_back(
         const_cast<float*>(conv_bias->template data<float>()),
         xft::xftVec<float>::dim_t{conv_bias->dims()[0]});
-#else
-    arg_conv_bias_.push_back(conv_bias->template data<float>());
-#endif
   }
 
   arg_fc_weight_int16_ = PrepareWeight<int16_t>(param.fc_weight);
   arg_conv_filter_int16_ = PrepareWeight<int16_t>(param.conv_weight);
   const int XPU_QUANT_SCALE_NUM = ctx.GetRawContext()->max_ptr_size();
   PrepareWeightMax(param.weight_max, XPU_QUANT_SCALE_NUM, &fc_weight_max_);
   PrepareFilterMax(param.conv_max, XPU_QUANT_SCALE_NUM, &conv_filter_max_);
-#ifdef USE_XFT
+
   int channel = static_cast<int>(param.input->dims()[1]);
   int xh = static_cast<int>(param.input->dims()[2]);
   int xw = static_cast<int>(param.input->dims()[3]);
@@ -239,7 +212,6 @@ void XPUSpatialTransformerCompute<InType, PType>::PrepareForRun() {
   st_param.strides = param.strides;
   st_param.gn_groups.push_back(param.groups);
   st_param.gn_eps.push_back(param.epsilon);
-#endif
 }
 
 template <typename InType, PrecisionType PType>
@@ -256,7 +228,6 @@ void XPUSpatialTransformerCompute<InType, PType>::Run() {
   int xw = static_cast<int>(param.input->dims()[3]);
   int embedding_seq = static_cast<int>(param.embedding->dims()[1]);
   int embedding_dim = static_cast<int>(param.embedding->dims()[2]);
-#ifdef USE_XFT
   // input
   xft::xftTensor<InType, 4> in_tensor(
       const_cast<InType*>(in), nullptr, {batch, channel, xh, xw});
@@ -286,40 +257,6 @@ void XPUSpatialTransformerCompute<InType, PType>::Run() {
       &output_tensor,
       st_param);
   CHECK_EQ(r, 0);
-#else
-  int r = xdnn::spatial_transformer_fusion<InType, int16_t, InType, int16_t>(
-      ctx.GetRawContext(),
-      in,
-      embedding,
-      *(XPUSpatialTransformerCompute::get_weight<int16_t>()),
-      *(XPUSpatialTransformerCompute::get_filter<int16_t>()),
-      out,
-      arg_fc_bias_,
-      arg_conv_bias_,
-      arg_ln_scale_,
-      arg_ln_bias_,
-      arg_gn_scale_,
-      arg_gn_bias_,
-      fc_weight_max_,
-      conv_filter_max_,
-      param.filter_dims,
-      param.dilations,
-      param.paddings,
-      param.strides,
-      param.conv_groups,
-      batch,
-      param.head_num,
-      param.size_per_head,
-      xh,
-      xw,
-      hidden_dim,
-      embedding_seq,
-      param.embedding_dim,
-      param.groups,
-      param.epsilon,
-      param.geglu_dim);
-  CHECK_EQ(r, 0);
-#endif
 }
 
 }  // namespace xpu

diff --git a/lite/operators/__xpu__spatial_transformer_op.cc b/lite/operators/__xpu__spatial_transformer_op.cc
@@ -40,11 +40,6 @@ bool XPUSpatialTransformerOp::CheckShape() const {
 }
 
 bool XPUSpatialTransformerOp::InferShapeImpl() const {
-  // auto input_shape = param_.input->dims();
-  // auto batch_size = input_shape[0];
-  // auto channel = input_shape[1];
-  // auto height = input_shape[2];
-  // auto weight = input_shape[3];
   param_.output->Resize(param_.input->dims());
   return true;
 }

diff --git a/lite/operators/__xpu__spatial_transformer_op.h b/lite/operators/__xpu__spatial_transformer_op.h
@@ -1,4 +1,4 @@
-// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+// Copyright (c) 2023 PaddlePaddle Authors. All Rights Reserved.
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.