Fix fused_attention_op and fused_feedforward_op bug when pre_layer_norm is false. (#36793)

* Fix bug when pre_layer_norm is false.

paddle/fluid/operators/fused/fused_attention_op.cc

@@ -37,12 +37,15 @@ class FusedAttentionOp : public framework::OperatorWithKernel {
     OP_INOUT_CHECK(ctx->HasInput("OutLinearBias"), "Input", "OutLinearBias",
                    "FusedAttentionOp");
 
-    OP_INOUT_CHECK(ctx->HasOutput("LnMean"), "Output", "LnMean",
-                   "FusedAttentionOp");
-    OP_INOUT_CHECK(ctx->HasOutput("LnVariance"), "Output", "LnVariance",
-                   "FusedAttentionOp");
-    OP_INOUT_CHECK(ctx->HasOutput("LnOut"), "Output", "LnOut",
-                   "FusedAttentionOp");
+    if (ctx->Attrs().Get<bool>("pre_layer_norm") == true) {
+      OP_INOUT_CHECK(ctx->HasOutput("LnMean"), "Output", "LnMean",
+                     "FusedAttentionOp");
+      OP_INOUT_CHECK(ctx->HasOutput("LnVariance"), "Output", "LnVariance",
+                     "FusedAttentionOp");
+      OP_INOUT_CHECK(ctx->HasOutput("LnOut"), "Output", "LnOut",
+                     "FusedAttentionOp");
+    }
+
     // qkv_out: [batch_size, seq_len, 3, num_head, dim_head]
     OP_INOUT_CHECK(ctx->HasOutput("QKVOut"), "Output", "QKVOut",
                    "FusedAttentionOp");
@@ -101,9 +104,11 @@ class FusedAttentionOp : public framework::OperatorWithKernel {
                           "input qkv_weight = [%s]",
                           x_dim, y_dim));
 
-    ctx->SetOutputDim("LnMean", {x_dim[0] * x_dim[1]});
-    ctx->SetOutputDim("LnVariance", {x_dim[0] * x_dim[1]});
-    ctx->SetOutputDim("LnOut", ctx->GetInputDim("X"));
+    if (ctx->Attrs().Get<bool>("pre_layer_norm") == true) {
+      ctx->SetOutputDim("LnMean", {x_dim[0] * x_dim[1]});
+      ctx->SetOutputDim("LnVariance", {x_dim[0] * x_dim[1]});
+      ctx->SetOutputDim("LnOut", ctx->GetInputDim("X"));
+    }
     // [batch_size, seq_len, 3, num_head, head_size]
     ctx->SetOutputDim("QKVOut",
                       {x_dim[0], x_dim[1], y_dim[0], y_dim[1], y_dim[2]});
@@ -351,11 +356,11 @@ class FusedAttentionGradOp : public framework::OperatorWithKernel {
                         ctx->GetInputDim("Ln2Bias"));
     }
     OP_INOUT_CHECK(ctx->HasInput("X"), "Input", "X", "FusedAttentionGrad");
-    OP_INOUT_CHECK(ctx->HasInput("LnMean"), "Input", "LnMean",
-                   "FusedAttentionGrad");
-    OP_INOUT_CHECK(ctx->HasInput("LnVariance"), "Input", "LnVariance",
-                   "FusedAttentionGrad");
     if (ctx->Attrs().Get<bool>("pre_layer_norm") == true) {
+      OP_INOUT_CHECK(ctx->HasInput("LnMean"), "Input", "LnMean",
+                     "FusedAttentionGrad");
+      OP_INOUT_CHECK(ctx->HasInput("LnVariance"), "Input", "LnVariance",
+                     "FusedAttentionGrad");
       OP_INOUT_CHECK(ctx->HasInput("LnOut"), "Input", "LnOut",
                      "FusedAttentionGrad");
     }
@@ -370,13 +375,15 @@ class FusedAttentionGradOp : public framework::OperatorWithKernel {
     OP_INOUT_CHECK(ctx->HasInput("OutLinearBias"), "Input", "OutLinearBias",
                    "FusedAttentionGrad");
 
-    if (ctx->HasOutput(framework::GradVarName("LnScale"))) {
-      ctx->SetOutputDim(framework::GradVarName("LnScale"),
-                        ctx->GetInputDim("LnScale"));
-    }
-    if (ctx->HasOutput(framework::GradVarName("LnBias"))) {
-      ctx->SetOutputDim(framework::GradVarName("LnBias"),
-                        ctx->GetInputDim("LnBias"));
+    if (ctx->Attrs().Get<bool>("pre_layer_norm") == true) {
+      if (ctx->HasOutput(framework::GradVarName("LnScale"))) {
+        ctx->SetOutputDim(framework::GradVarName("LnScale"),
+                          ctx->GetInputDim("LnScale"));
+      }
+      if (ctx->HasOutput(framework::GradVarName("LnBias"))) {
+        ctx->SetOutputDim(framework::GradVarName("LnBias"),
+                          ctx->GetInputDim("LnBias"));
+      }
     }
     if (ctx->HasOutput(framework::GradVarName("X"))) {
       ctx->SetOutputDim(framework::GradVarName("X"), ctx->GetInputDim("X"));
@@ -390,8 +397,10 @@ class FusedAttentionGradOp : public framework::OperatorWithKernel {
     ctx->SetOutputDim(framework::GradVarName("QKVBias"),
                       ctx->GetInputDim("QKVBias"));
 
-    ctx->SetOutputDim(framework::GradVarName("LnOut"),
-                      ctx->GetInputDim("LnOut"));
+    if (ctx->Attrs().Get<bool>("pre_layer_norm") == true) {
+      ctx->SetOutputDim(framework::GradVarName("LnOut"),
+                        ctx->GetInputDim("LnOut"));
+    }
     ctx->SetOutputDim(framework::GradVarName("FMHAOut"),
                       ctx->GetInputDim("FMHAOut"));
     ctx->SetOutputDim(framework::GradVarName("QKTVOut"),
@@ -442,16 +451,23 @@ class FusedAttentionGradOpMaker : public framework::SingleGradOpMaker<T> {
     op->SetInput("SrcMask", this->Input("SrcMask"));
     op->SetInput("OutLinearW", this->Input("OutLinearW"));
     op->SetInput("OutLinearBias", this->Input("OutLinearBias"));
-    if (this->HasInput("LnScale")) {
-      op->SetInput("LnScale", this->Input("LnScale"));
-      op->SetOutput(framework::GradVarName("LnScale"),
-                    this->InputGrad("LnScale"));
-    }
-    if (this->HasInput("LnBias")) {
-      op->SetInput("LnBias", this->Input("LnBias"));
-      op->SetOutput(framework::GradVarName("LnBias"),
-                    this->InputGrad("LnBias"));
+
+    op->SetAttrMap(this->Attrs());
+    bool is_pre_layer_norm =
+        BOOST_GET_CONST(bool, op->GetAttr("pre_layer_norm"));
+    if (is_pre_layer_norm) {
+      if (this->HasInput("LnScale")) {
+        op->SetInput("LnScale", this->Input("LnScale"));
+        op->SetOutput(framework::GradVarName("LnScale"),
+                      this->InputGrad("LnScale"));
+      }
+      if (this->HasInput("LnBias")) {
+        op->SetInput("LnBias", this->Input("LnBias"));
+        op->SetOutput(framework::GradVarName("LnBias"),
+                      this->InputGrad("LnBias"));
+      }
     }
+
     if (this->HasInput("Ln2Scale")) {
       op->SetInput("Ln2Scale", this->Input("Ln2Scale"));
       op->SetOutput(framework::GradVarName("Ln2Scale"),
@@ -473,9 +489,17 @@ class FusedAttentionGradOpMaker : public framework::SingleGradOpMaker<T> {
                   this->InputGrad("OutLinearW"));
 
     // use forward outputs as backward inputs.
-    op->SetInput("LnOut", this->Output("LnOut"));
-    op->SetInput("LnMean", this->Output("LnMean"));
-    op->SetInput("LnVariance", this->Output("LnVariance"));
+    if (is_pre_layer_norm) {
+      if (this->HasOutput("LnOut")) {
+        op->SetInput("LnOut", this->Output("LnOut"));
+      }
+      if (this->HasOutput("LnMean")) {
+        op->SetInput("LnMean", this->Output("LnMean"));
+      }
+      if (this->HasOutput("LnVariance")) {
+        op->SetInput("LnVariance", this->Output("LnVariance"));
+      }
+    }
     op->SetInput("QKVOut", this->Output("QKVOut"));
     op->SetInput("QKVBiasOut", this->Output("QKVBiasOut"));
     op->SetInput("TransposeOut2", this->Output("TransposeOut2"));
@@ -496,7 +520,12 @@ class FusedAttentionGradOpMaker : public framework::SingleGradOpMaker<T> {
     op->SetInput("QKVOut", this->Output("QKVOut"));
 
     // backward outputs: dinput
-    op->SetOutput(framework::GradVarName("LnOut"), this->OutputGrad("LnOut"));
+    if (is_pre_layer_norm) {
+      if (this->HasOutput("LnOut")) {
+        op->SetOutput(framework::GradVarName("LnOut"),
+                      this->OutputGrad("LnOut"));
+      }
+    }
     op->SetOutput(framework::GradVarName("QKVOut"), this->OutputGrad("QKVOut"));
     op->SetOutput(framework::GradVarName("QKVBiasOut"),
                   this->OutputGrad("QKVBiasOut"));
@@ -517,8 +546,6 @@ class FusedAttentionGradOpMaker : public framework::SingleGradOpMaker<T> {
                   this->OutputGrad("BiasDropoutResidualOut"));
     op->SetOutput(framework::GradVarName("OutLinearOut"),
                   this->OutputGrad("OutLinearOut"));
-
-    op->SetAttrMap(this->Attrs());
   }
 };
 

paddle/fluid/operators/fused/fused_attention_op.cu

@@ -97,9 +97,12 @@ class FusedAttentionOpKernel : public framework::OpKernel<T> {
     auto *x_data = input_x->data<T>();
     auto *ln_scale_data = (ln_scale == nullptr ? nullptr : ln_scale->data<U>());
     auto *ln_bias_data = (ln_bias == nullptr ? nullptr : ln_bias->data<U>());
-    auto *ln_mean_data = ln_mean->mutable_data<U>(ctx.GetPlace());
-    auto *ln_var_data = ln_var->mutable_data<U>(ctx.GetPlace());
-    auto *ln_out_data = ln_out->mutable_data<T>(ctx.GetPlace());
+    auto *ln_mean_data =
+        pre_layer_norm ? ln_mean->mutable_data<U>(ctx.GetPlace()) : nullptr;
+    auto *ln_var_data =
+        pre_layer_norm ? ln_var->mutable_data<U>(ctx.GetPlace()) : nullptr;
+    auto *ln_out_data =
+        pre_layer_norm ? ln_out->mutable_data<T>(ctx.GetPlace()) : nullptr;
 
     auto *qkv_weight_data = qkv_weight->data<T>();
     auto *qkv_bias_data = qkv_bias->data<T>();
@@ -243,9 +246,6 @@ class FusedAttentionGradKernel : public framework::OpKernel<T> {
     auto *out_linear_bias_data = out_linear_bias->data<T>();
 
     // fw output
-    auto *ln_mean = ctx.Input<Tensor>("LnMean");
-    auto *ln_var = ctx.Input<Tensor>("LnVariance");
-    auto *ln_out = ctx.Input<Tensor>("LnOut");
     auto *fmha_out = ctx.Input<Tensor>("FMHAOut");
     auto *transpose_out_2 = ctx.Input<Tensor>("TransposeOut2");
     auto *qk_out = ctx.Input<Tensor>("QKOut");
@@ -260,9 +260,6 @@ class FusedAttentionGradKernel : public framework::OpKernel<T> {
     auto *dropout_mask_out = ctx.Input<Tensor>("DropoutMaskOut");
     auto *bias_dropout_residual_out =
         ctx.Input<Tensor>("BiasDropoutResidualOut");
-    auto *ln_mean_data = ln_mean->data<U>();
-    auto *ln_var_data = ln_var->data<U>();
-    auto *ln_out_data = ln_out->data<T>();
     auto *fmha_out_data = fmha_out->data<T>();
     auto *transpose_out_2_data = transpose_out_2->data<T>();
     auto *qk_out_data = qk_out->data<T>();
@@ -277,7 +274,6 @@ class FusedAttentionGradKernel : public framework::OpKernel<T> {
 
     // output's grad
     auto *d_x = ctx.Output<Tensor>(framework::GradVarName("X"));
-    auto *d_ln_out = ctx.Output<Tensor>(framework::GradVarName("LnOut"));
     auto *d_qkv_out = ctx.Output<Tensor>(framework::GradVarName("QKVOut"));
     auto *d_qkv_bias_out =
         ctx.Output<Tensor>(framework::GradVarName("QKVBiasOut"));
@@ -297,7 +293,6 @@ class FusedAttentionGradKernel : public framework::OpKernel<T> {
     auto *d_bias_dropout_residual_out =
         ctx.Output<Tensor>(framework::GradVarName("BiasDropoutResidualOut"));
     auto *d_x_data = d_x->mutable_data<T>(ctx.GetPlace());
-    auto *d_ln_out_data = d_ln_out->mutable_data<T>(ctx.GetPlace());
     auto *d_qkv_out_data = d_qkv_out->mutable_data<T>(ctx.GetPlace());
     auto *d_qkv_bias_out_data = d_qkv_bias_out->mutable_data<T>(ctx.GetPlace());
     auto *d_qktv_out_data = d_qktv_out->mutable_data<T>(ctx.GetPlace());
@@ -315,8 +310,6 @@ class FusedAttentionGradKernel : public framework::OpKernel<T> {
         d_bias_dropout_residual_out->mutable_data<T>(ctx.GetPlace());
 
     // parameter grad
-    auto *d_ln_scale = ctx.Output<Tensor>(framework::GradVarName("LnScale"));
-    auto *d_ln_bias = ctx.Output<Tensor>(framework::GradVarName("LnBias"));
     auto *d_qkv_weight = ctx.Output<Tensor>(framework::GradVarName("QKVW"));
     auto *d_qkv_bias = ctx.Output<Tensor>(framework::GradVarName("QKVBias"));
     auto *d_out_linear_weight =
@@ -325,12 +318,7 @@ class FusedAttentionGradKernel : public framework::OpKernel<T> {
         ctx.Output<Tensor>(framework::GradVarName("OutLinearBias"));
     auto *d_ln_2_scale = ctx.Output<Tensor>(framework::GradVarName("Ln2Scale"));
     auto *d_ln_2_bias = ctx.Output<Tensor>(framework::GradVarName("Ln2Bias"));
-    auto *d_ln_scale_data =
-        (d_ln_scale == nullptr ? nullptr
-                               : d_ln_scale->mutable_data<U>(ctx.GetPlace()));
-    auto *d_ln_bias_data =
-        (d_ln_bias == nullptr ? nullptr
-                              : d_ln_bias->mutable_data<U>(ctx.GetPlace()));
+
     auto *d_qkv_weight_data = d_qkv_weight->mutable_data<T>(ctx.GetPlace());
     auto *d_qkv_bias_data = d_qkv_bias->mutable_data<T>(ctx.GetPlace());
     auto *d_out_linear_weight_data =
@@ -407,6 +395,24 @@ class FusedAttentionGradKernel : public framework::OpKernel<T> {
                     cudaMemcpyDeviceToDevice);
 
     if (pre_layer_norm) {
+      auto *ln_mean = ctx.Input<Tensor>("LnMean");
+      auto *ln_var = ctx.Input<Tensor>("LnVariance");
+      auto *ln_out = ctx.Input<Tensor>("LnOut");
+      auto *ln_mean_data = ln_mean->data<U>();
+      auto *ln_var_data = ln_var->data<U>();
+      auto *ln_out_data = ln_out->data<T>();
+
+      auto *d_ln_out = ctx.Output<Tensor>(framework::GradVarName("LnOut"));
+      auto *d_ln_scale = ctx.Output<Tensor>(framework::GradVarName("LnScale"));
+      auto *d_ln_bias = ctx.Output<Tensor>(framework::GradVarName("LnBias"));
+      auto *d_ln_out_data = d_ln_out->mutable_data<T>(ctx.GetPlace());
+      auto *d_ln_scale_data =
+          (d_ln_scale == nullptr ? nullptr
+                                 : d_ln_scale->mutable_data<U>(ctx.GetPlace()));
+      auto *d_ln_bias_data =
+          (d_ln_bias == nullptr ? nullptr
+                                : d_ln_bias->mutable_data<U>(ctx.GetPlace()));
+
       qkv_compute.ComputeBackward(ln_out_data, qkv_weight_data,
                                   d_qkv_bias_out_data, d_ln_out_data,
                                   d_qkv_weight_data, d_qkv_bias_data);

paddle/fluid/operators/fused/fused_feedforward_op.cc

@@ -41,18 +41,8 @@ class FusedFeedForwardOp : public framework::OperatorWithKernel {
                    "fused_feedforward");
     OP_INOUT_CHECK(context->HasOutput("Dropout2Mask"), "Output", "Dropout2Mask",
                    "fused_feedforward");
-    OP_INOUT_CHECK(context->HasOutput("Ln1Mean"), "Output", "Ln1Mean",
-                   "fused_feedforward");
-    OP_INOUT_CHECK(context->HasOutput("Ln1Variance"), "Output", "Ln1Variance",
-                   "fused_feedforward");
-    OP_INOUT_CHECK(context->HasOutput("Ln2Mean"), "Output", "Ln2Mean",
-                   "fused_feedforward");
-    OP_INOUT_CHECK(context->HasOutput("Ln2Variance"), "Output", "Ln2Variance",
-                   "fused_feedforward");
     OP_INOUT_CHECK(context->HasOutput("Linear1Out"), "Output", "Linear1Out",
                    "fused_feedforward");
-    OP_INOUT_CHECK(context->HasOutput("Ln1Out"), "Output", "Ln1Out",
-                   "fused_feedforward");
     OP_INOUT_CHECK(context->HasOutput("Dropout1Out"), "Output", "Dropout1Out",
                    "fused_feedforward");
     OP_INOUT_CHECK(context->HasOutput("Dropout2Out"), "Output", "Dropout2Out",
@@ -76,7 +66,6 @@ class FusedFeedForwardOp : public framework::OperatorWithKernel {
     }
     context->SetOutputDim("Dropout1Out", tmp_dim_x);
     context->SetOutputDim("Linear1Out", tmp_dim_x);
-    context->SetOutputDim("Ln1Out", dim_x);
     context->SetOutputDim("Dropout2Out", dim_x);
 
     if (context->Attrs().Get<bool>("dropout2_is_test") == false) {
@@ -84,10 +73,25 @@ class FusedFeedForwardOp : public framework::OperatorWithKernel {
     }
     framework::DDim mean_dim =
         framework::make_ddim({mat_dim_x.batch_size_ * mat_dim_x.height_});
-    context->SetOutputDim("Ln1Mean", mean_dim);
-    context->SetOutputDim("Ln1Variance", mean_dim);
-    context->SetOutputDim("Ln2Mean", mean_dim);
-    context->SetOutputDim("Ln2Variance", mean_dim);
+    bool pre_layer_norm = context->Attrs().Get<bool>("pre_layer_norm");
+    if (pre_layer_norm) {
+      OP_INOUT_CHECK(context->HasOutput("Ln1Mean"), "Output", "Ln1Mean",
+                     "fused_feedforward");
+      OP_INOUT_CHECK(context->HasOutput("Ln1Variance"), "Output", "Ln1Variance",
+                     "fused_feedforward");
+      OP_INOUT_CHECK(context->HasOutput("Ln1Out"), "Output", "Ln1Out",
+                     "fused_feedforward");
+      context->SetOutputDim("Ln1Out", dim_x);
+      context->SetOutputDim("Ln1Mean", mean_dim);
+      context->SetOutputDim("Ln1Variance", mean_dim);
+    } else {
+      OP_INOUT_CHECK(context->HasOutput("Ln2Mean"), "Output", "Ln2Mean",
+                     "fused_feedforward");
+      OP_INOUT_CHECK(context->HasOutput("Ln2Variance"), "Output", "Ln2Variance",
+                     "fused_feedforward");
+      context->SetOutputDim("Ln2Mean", mean_dim);
+      context->SetOutputDim("Ln2Variance", mean_dim);
+    }
     context->ShareLoD("X", "Out");
   }
 
@@ -218,14 +222,13 @@ class FusedFeedForwardOpGrad : public framework::OperatorWithKernel {
     PADDLE_ENFORCE_EQ(ctx->Attrs().Get<bool>("dropout2_is_test"), false,
                       platform::errors::InvalidArgument(
                           "GradOp is only callable when is_test is false"));
+    bool pre_layer_norm = ctx->Attrs().Get<bool>("pre_layer_norm");
     OP_INOUT_CHECK(ctx->HasInput("Dropout1Mask"), "Input", "Dropout1Mask",
                    "FusedFeedForwardGrad");
     OP_INOUT_CHECK(ctx->HasInput("Dropout2Mask"), "Input", "Dropout1Mask",
                    "FusedFeedForwardGrad");
     OP_INOUT_CHECK(ctx->HasInput("Linear1Out"), "Input", "Linear1Out",
                    "FusedFeedForwardGrad");
-    OP_INOUT_CHECK(ctx->HasInput("Ln1Out"), "Input", "Ln1Out",
-                   "FusedFeedForwardGrad");
     OP_INOUT_CHECK(ctx->HasInput("Dropout1Out"), "Input", "Dropout1Out",
                    "FusedFeedForwardGrad");
     OP_INOUT_CHECK(ctx->HasInput("Dropout2Out"), "Input", "Dropout2Out",
@@ -234,14 +237,19 @@ class FusedFeedForwardOpGrad : public framework::OperatorWithKernel {
                    "FusedFeedForwardGrad");
     OP_INOUT_CHECK(ctx->HasInput("Linear2Weight"), "Input", "Linear2Weight",
                    "FusedFeedForwardGrad");
-    OP_INOUT_CHECK(ctx->HasInput("Ln1Mean"), "Input", "Ln1Mean",
-                   "FusedFeedForwardGrad");
-    OP_INOUT_CHECK(ctx->HasInput("Ln1Variance"), "Input", "Ln1Variance",
-                   "FusedFeedForwardGrad");
-    OP_INOUT_CHECK(ctx->HasInput("Ln2Mean"), "Input", "Ln2Mean",
-                   "FusedFeedForwardGrad");
-    OP_INOUT_CHECK(ctx->HasInput("Ln2Variance"), "Input", "Ln2Variance",
-                   "FusedFeedForwardGrad");
+    if (pre_layer_norm) {
+      OP_INOUT_CHECK(ctx->HasInput("Ln1Mean"), "Input", "Ln1Mean",
+                     "FusedFeedForwardGrad");
+      OP_INOUT_CHECK(ctx->HasInput("Ln1Variance"), "Input", "Ln1Variance",
+                     "FusedFeedForwardGrad");
+      OP_INOUT_CHECK(ctx->HasInput("Ln1Out"), "Input", "Ln1Out",
+                     "FusedFeedForwardGrad");
+    } else {
+      OP_INOUT_CHECK(ctx->HasInput("Ln2Mean"), "Input", "Ln2Mean",
+                     "FusedFeedForwardGrad");
+      OP_INOUT_CHECK(ctx->HasInput("Ln2Variance"), "Input", "Ln2Variance",
+                     "FusedFeedForwardGrad");
+    }
 
     OP_INOUT_CHECK(ctx->HasInput(framework::GradVarName("Out")), "Input",
                    framework::GradVarName("Out"), "FusedFeedForwardGrad");
@@ -299,30 +307,36 @@ class FusedFeedForwardOpGradMaker : public framework::SingleGradOpMaker<T> {
     op->SetInput("Linear1Weight", this->Input("Linear1Weight"));
     op->SetInput("Linear1Bias", this->Input("Linear1Bias"));
     op->SetInput("Linear2Weight", this->Input("Linear2Weight"));
-    op->SetInput("Ln1Scale", this->Input("Ln1Scale"));
-    op->SetInput("Ln1Bias", this->Input("Ln1Bias"));
-    op->SetInput("Ln2Scale", this->Input("Ln2Scale"));
-    op->SetInput("Ln2Bias", this->Input("Ln2Bias"));
     op->SetInput("Dropout1Mask", this->Output("Dropout1Mask"));
     op->SetInput("Dropout2Mask", this->Output("Dropout2Mask"));
     op->SetInput("Linear1Out", this->Output("Linear1Out"));
-    op->SetInput("Ln1Out", this->Output("Ln1Out"));
-    op->SetInput("Ln1Mean", this->Output("Ln1Mean"));
-    op->SetInput("Ln1Variance", this->Output("Ln1Variance"));
-    op->SetInput("Ln2Mean", this->Output("Ln2Mean"));
-    op->SetInput("Ln2Variance", this->Output("Ln2Variance"));
     op->SetInput("Dropout1Out", this->Output("Dropout1Out"));
     op->SetInput("Dropout2Out", this->Output("Dropout2Out"));
 
+    op->SetAttrMap(this->Attrs());
+    bool pre_layer_norm = BOOST_GET_CONST(bool, op->GetAttr("pre_layer_norm"));
+
     op->SetOutput(framework::GradVarName("X"), this->InputGrad("X"));
-    op->SetOutput(framework::GradVarName("Ln1Scale"),
-                  this->InputGrad("Ln1Scale"));
-    op->SetOutput(framework::GradVarName("Ln1Bias"),
-                  this->InputGrad("Ln1Bias"));
-    op->SetOutput(framework::GradVarName("Ln2Scale"),
-                  this->InputGrad("Ln2Scale"));
-    op->SetOutput(framework::GradVarName("Ln2Bias"),
-                  this->InputGrad("Ln2Bias"));
+    if (pre_layer_norm) {
+      op->SetInput("Ln1Scale", this->Input("Ln1Scale"));
+      op->SetInput("Ln1Bias", this->Input("Ln1Bias"));
+      op->SetInput("Ln1Out", this->Output("Ln1Out"));
+      op->SetInput("Ln1Mean", this->Output("Ln1Mean"));
+      op->SetInput("Ln1Variance", this->Output("Ln1Variance"));
+      op->SetOutput(framework::GradVarName("Ln1Scale"),
+                    this->InputGrad("Ln1Scale"));
+      op->SetOutput(framework::GradVarName("Ln1Bias"),
+                    this->InputGrad("Ln1Bias"));
+    } else {
+      op->SetInput("Ln2Scale", this->Input("Ln2Scale"));
+      op->SetInput("Ln2Bias", this->Input("Ln2Bias"));
+      op->SetInput("Ln2Mean", this->Output("Ln2Mean"));
+      op->SetInput("Ln2Variance", this->Output("Ln2Variance"));
+      op->SetOutput(framework::GradVarName("Ln2Scale"),
+                    this->InputGrad("Ln2Scale"));
+      op->SetOutput(framework::GradVarName("Ln2Bias"),
+                    this->InputGrad("Ln2Bias"));
+    }
     op->SetOutput(framework::GradVarName("Linear1Weight"),
                   this->InputGrad("Linear1Weight"));
     op->SetOutput(framework::GradVarName("Linear1Bias"),
@@ -334,8 +348,6 @@ class FusedFeedForwardOpGradMaker : public framework::SingleGradOpMaker<T> {
       op->SetOutput(framework::GradVarName("Linear2Bias"),
                     this->InputGrad("Linear2Bias"));
     }
-
-    op->SetAttrMap(this->Attrs());
   }
 };
 

paddle/fluid/operators/fused/fused_feedforward_op.cu

@@ -113,26 +113,40 @@ class FusedFeedForwardKernel : public framework::OpKernel<T> {
     auto* linear1_bias = context.Input<framework::Tensor>("Linear1Bias");
     auto* linear2_weight = context.Input<framework::Tensor>("Linear2Weight");
     auto* linear2_bias = context.Input<framework::Tensor>("Linear2Bias");
-    auto* ln1_scale = context.Input<framework::Tensor>("Ln1Scale");
-    auto* ln1_bias = context.Input<framework::Tensor>("Ln1Bias");
-    auto* ln2_scale = context.Input<framework::Tensor>("Ln2Scale");
-    auto* ln2_bias = context.Input<framework::Tensor>("Ln2Bias");
-
-    auto* ln1_mean = context.Output<framework::Tensor>("Ln1Mean");
-    auto* ln1_variance = context.Output<framework::Tensor>("Ln1Variance");
-    auto* ln2_mean = context.Output<framework::Tensor>("Ln2Mean");
-    auto* ln2_variance = context.Output<framework::Tensor>("Ln2Variance");
+    const bool pre_layer_norm = context.Attr<bool>("pre_layer_norm");
+
+    auto* ln1_scale =
+        pre_layer_norm ? context.Input<framework::Tensor>("Ln1Scale") : nullptr;
+    auto* ln1_bias =
+        pre_layer_norm ? context.Input<framework::Tensor>("Ln1Bias") : nullptr;
+    auto* ln2_scale = !pre_layer_norm
+                          ? context.Input<framework::Tensor>("Ln2Scale")
+                          : nullptr;
+    auto* ln2_bias =
+        !pre_layer_norm ? context.Input<framework::Tensor>("Ln2Bias") : nullptr;
+
+    auto* ln1_mean =
+        pre_layer_norm ? context.Output<framework::Tensor>("Ln1Mean") : nullptr;
+    auto* ln1_variance = pre_layer_norm
+                             ? context.Output<framework::Tensor>("Ln1Variance")
+                             : nullptr;
+    auto* ln2_mean = !pre_layer_norm
+                         ? context.Output<framework::Tensor>("Ln2Mean")
+                         : nullptr;
+    auto* ln2_variance = !pre_layer_norm
+                             ? context.Output<framework::Tensor>("Ln2Variance")
+                             : nullptr;
     auto* out = context.Output<framework::Tensor>("Out");
     auto* dropout1_mask = context.Output<framework::Tensor>("Dropout1Mask");
     auto* dropout2_mask = context.Output<framework::Tensor>("Dropout2Mask");
     auto* linear1_out = context.Output<framework::Tensor>("Linear1Out");
-    auto* ln1_out = context.Output<framework::Tensor>("Ln1Out");
+    auto* ln1_out =
+        pre_layer_norm ? context.Output<framework::Tensor>("Ln1Out") : nullptr;
     auto* dropout1_out = context.Output<framework::Tensor>("Dropout1Out");
     auto* dropout2_out = context.Output<framework::Tensor>("Dropout2Out");
 
     const std::string act_method = context.Attr<std::string>("act_method");
 
-    const bool pre_layer_norm = context.Attr<bool>("pre_layer_norm");
     const float epsilon1 = context.Attr<float>("ln1_epsilon");
     const float epsilon2 = context.Attr<float>("ln2_epsilon");
 
@@ -144,12 +158,16 @@ class FusedFeedForwardKernel : public framework::OpKernel<T> {
     out->mutable_data<T>(place);
     dropout1_mask->mutable_data<uint8_t>(place);
     dropout2_mask->mutable_data<uint8_t>(place);
-    ln1_mean->mutable_data<U>(place);
-    ln1_variance->mutable_data<U>(place);
-    ln2_mean->mutable_data<U>(place);
-    ln2_variance->mutable_data<U>(place);
+    if (pre_layer_norm) {
+      ln1_mean->mutable_data<U>(place);
+      ln1_variance->mutable_data<U>(place);
+      ln1_out->mutable_data<T>(place);
+    } else {
+      ln2_mean->mutable_data<U>(place);
+      ln2_variance->mutable_data<U>(place);
+    }
+
     linear1_out->mutable_data<T>(place);
-    ln1_out->mutable_data<T>(place);
     dropout1_out->mutable_data<T>(place);
     dropout2_out->mutable_data<T>(place);
 
@@ -193,16 +211,16 @@ class FusedFeedForwardGradKernel : public framework::OpKernel<T> {
       const framework::Tensor& d_out, const framework::Tensor& x,
       const framework::Tensor& dropout1_mask,
       const framework::Tensor& dropout2_mask,
-      const framework::Tensor& linear1_out, const framework::Tensor& ln1_out,
+      const framework::Tensor& linear1_out, const framework::Tensor* ln1_out,
       const framework::Tensor& dropout1_out,
       const framework::Tensor& dropout2_out,
       const framework::Tensor& linear1_weight,
       const framework::Tensor* linear1_bias,
       const framework::Tensor& linear2_weight,
       const framework::Tensor* ln1_gamma, const framework::Tensor* ln1_beta,
-      const framework::Tensor& ln1_mean, const framework::Tensor& ln1_variance,
+      const framework::Tensor* ln1_mean, const framework::Tensor* ln1_variance,
       const framework::Tensor* ln2_gamma, const framework::Tensor* ln2_beta,
-      const framework::Tensor& ln2_mean, const framework::Tensor& ln2_variance,
+      const framework::Tensor* ln2_mean, const framework::Tensor* ln2_variance,
       framework::Tensor* d_x, framework::Tensor* d_linear1_weight,
       framework::Tensor* d_linear1_bias, framework::Tensor* d_linear2_weight,
       framework::Tensor* d_linear2_bias, framework::Tensor* d_ln1_gamma,
@@ -252,8 +270,8 @@ class FusedFeedForwardGradKernel : public framework::OpKernel<T> {
     } else {
       fused_dropout_layernorm_helper.LayernormResidualDropoutBiasGrad(
           ctx, d_out.data<T>(), dropout2_out.data<T>(),
-          dropout2_mask.data<uint8_t>(), ln2_gamma_ptr, ln2_mean.data<U>(),
-          ln2_variance.data<U>(), d_dropout2_out.data<T>(), d_ln2_gamma_ptr,
+          dropout2_mask.data<uint8_t>(), ln2_gamma_ptr, ln2_mean->data<U>(),
+          ln2_variance->data<U>(), d_dropout2_out.data<T>(), d_ln2_gamma_ptr,
           d_ln2_beta_ptr, d_linear2_out.data<T>(), d_linear2_bias_ptr,
           d_residual.data<T>());
     }
@@ -273,13 +291,13 @@ class FusedFeedForwardGradKernel : public framework::OpKernel<T> {
     if (pre_layer_norm) {
       framework::Tensor d_ln1_out;
       d_ln1_out.mutable_data<T>({bsz_seq, d_model}, place);
-      MatMulGrad(ctx, d_linear1_out, ln1_out, linear1_weight, &d_ln1_out,
+      MatMulGrad(ctx, d_linear1_out, *ln1_out, linear1_weight, &d_ln1_out,
                  d_linear1_weight);
 
-      pre_layernorm_helper.LayerNormGrad(ctx, d_ln1_out.data<T>(), x.data<T>(),
-                                         ln1_gamma_ptr, ln1_mean.data<U>(),
-                                         ln1_variance.data<U>(), d_x->data<T>(),
-                                         d_ln1_gamma_ptr, d_ln1_beta_ptr);
+      pre_layernorm_helper.LayerNormGrad(
+          ctx, d_ln1_out.data<T>(), x.data<T>(), ln1_gamma_ptr,
+          ln1_mean->data<U>(), ln1_variance->data<U>(), d_x->data<T>(),
+          d_ln1_gamma_ptr, d_ln1_beta_ptr);
     } else {
       MatMulGrad(ctx, d_linear1_out, x, linear1_weight, d_x, d_linear1_weight);
     }
@@ -290,33 +308,52 @@ class FusedFeedForwardGradKernel : public framework::OpKernel<T> {
     auto d_out =
         *context.Input<framework::Tensor>(framework::GradVarName("Out"));
     auto x = *context.Input<framework::Tensor>("X");
+    const bool pre_layer_norm = context.Attr<bool>("pre_layer_norm");
     auto dropout1_mask = *context.Input<framework::Tensor>("Dropout1Mask");
     auto dropout2_mask = *context.Input<framework::Tensor>("Dropout2Mask");
     auto linear1_out = *context.Input<framework::Tensor>("Linear1Out");
-    auto ln1_out = *context.Input<framework::Tensor>("Ln1Out");
+    auto* ln1_out =
+        pre_layer_norm ? context.Input<framework::Tensor>("Ln1Out") : nullptr;
     auto dropout1_out = *context.Input<framework::Tensor>("Dropout1Out");
     auto dropout2_out = *context.Input<framework::Tensor>("Dropout2Out");
     auto linear1_weight = *context.Input<framework::Tensor>("Linear1Weight");
     auto* linear1_bias = context.Input<framework::Tensor>("Linear1Bias");
     auto linear2_weight = *context.Input<framework::Tensor>("Linear2Weight");
-    auto ln1_mean = *context.Input<framework::Tensor>("Ln1Mean");
-    auto ln1_variance = *context.Input<framework::Tensor>("Ln1Variance");
-    auto* ln1_scale = context.Input<framework::Tensor>("Ln1Scale");
-    auto* ln1_bias = context.Input<framework::Tensor>("Ln1Bias");
-    auto ln2_mean = *context.Input<framework::Tensor>("Ln2Mean");
-    auto ln2_variance = *context.Input<framework::Tensor>("Ln2Variance");
-    auto* ln2_scale = context.Input<framework::Tensor>("Ln2Scale");
-    auto* ln2_bias = context.Input<framework::Tensor>("Ln2Bias");
+    auto* ln1_mean =
+        pre_layer_norm ? context.Input<framework::Tensor>("Ln1Mean") : nullptr;
+    auto* ln1_variance = pre_layer_norm
+                             ? context.Input<framework::Tensor>("Ln1Variance")
+                             : nullptr;
+    auto* ln1_scale =
+        pre_layer_norm ? context.Input<framework::Tensor>("Ln1Scale") : nullptr;
+    auto* ln1_bias =
+        pre_layer_norm ? context.Input<framework::Tensor>("Ln1Bias") : nullptr;
+    auto* ln2_mean =
+        !pre_layer_norm ? context.Input<framework::Tensor>("Ln2Mean") : nullptr;
+    auto* ln2_variance = !pre_layer_norm
+                             ? context.Input<framework::Tensor>("Ln2Variance")
+                             : nullptr;
+    auto* ln2_scale = !pre_layer_norm
+                          ? context.Input<framework::Tensor>("Ln2Scale")
+                          : nullptr;
+    auto* ln2_bias =
+        !pre_layer_norm ? context.Input<framework::Tensor>("Ln2Bias") : nullptr;
 
     auto* d_x = context.Output<framework::Tensor>(framework::GradVarName("X"));
-    auto* d_ln1_scale =
-        context.Output<framework::Tensor>(framework::GradVarName("Ln1Scale"));
-    auto* d_ln1_bias =
-        context.Output<framework::Tensor>(framework::GradVarName("Ln1Bias"));
+    auto* d_ln1_scale = pre_layer_norm
+                            ? context.Output<framework::Tensor>(
+                                  framework::GradVarName("Ln1Scale"))
+                            : nullptr;
+    auto* d_ln1_bias = pre_layer_norm
+                           ? context.Output<framework::Tensor>(
+                                 framework::GradVarName("Ln1Bias"))
+                           : nullptr;
     auto* d_ln2_scale =
-        context.Output<framework::Tensor>(framework::GradVarName("Ln2Scale"));
+        pre_layer_norm ? nullptr : context.Output<framework::Tensor>(
+                                       framework::GradVarName("Ln2Scale"));
     auto* d_ln2_bias =
-        context.Output<framework::Tensor>(framework::GradVarName("Ln2Bias"));
+        pre_layer_norm ? nullptr : context.Output<framework::Tensor>(
+                                       framework::GradVarName("Ln2Bias"));
     auto* d_linear1_weight = context.Output<framework::Tensor>(
         framework::GradVarName("Linear1Weight"));
     auto* d_linear1_bias = context.Output<framework::Tensor>(
@@ -328,7 +365,6 @@ class FusedFeedForwardGradKernel : public framework::OpKernel<T> {
 
     const float epsilon1 = context.Attr<float>("ln1_epsilon");
     const float epsilon2 = context.Attr<float>("ln2_epsilon");
-    const bool pre_layer_norm = context.Attr<bool>("pre_layer_norm");
     const std::string act_method = context.Attr<std::string>("act_method");
     DropoutParam dropout_param1(context, 1);
     DropoutParam dropout_param2(context, 2);

python/paddle/fluid/tests/unittests/test_fused_attention_op.py

@@ -65,7 +65,7 @@ class TestFusedAttentionOp(OpTest):
     def config(self):
         self.x_type = np.float32
         self.attn_mask_type = np.float64
-        self.pre_layer_norm = True
+        self.pre_layer_norm = False
         self.training = True
 
         self.batch_size = 8
@@ -213,11 +213,40 @@ class TestFusedAttentionOp(OpTest):
             x_grad_ref, x_grad.numpy(), rtol=1e-5, atol=1e-5)
 
 
+class TestFusedAttentionOpPreLn(TestFusedAttentionOp):
+    def config(self):
+        self.x_type = np.float32
+        self.attn_mask_type = np.float64
+        self.pre_layer_norm = True
+        self.training = True
+
+        self.batch_size = 8
+        self.query_length = 128
+        self.head_dim = 64
+        self.num_heads = 16
+        self.embed_dim = self.head_dim * self.num_heads
+
+        self.dropout_prob = 0.0
+        self.attn_dropout_prob = 0.0
+        self.weight_attr = None
+        self.bias_attr = None
+        self.kdim, self.vdim = self.embed_dim, self.embed_dim
+        self.key_length, self.value_length = self.query_length, self.query_length
+
+    def test_fused_attention_op(self):
+        final_out_ref, x_grad_ref = self.GetBaselineOut()
+        final_out, x_grad = self.GetFusedAttentionOut()
+        np.testing.assert_allclose(
+            final_out_ref, final_out.numpy(), rtol=1e-5, atol=1e-1)
+        np.testing.assert_allclose(
+            x_grad_ref, x_grad.numpy(), rtol=1e-5, atol=1e-1)
+
+
 class TestFusedAttentionOpFp16(TestFusedAttentionOp):
     def config(self):
         self.x_type = np.float16
         self.attn_mask_type = np.float64
-        self.pre_layer_norm = True
+        self.pre_layer_norm = False
         self.training = True
 
         self.batch_size = 8