From c3d3001fa7fc0fddae07a0870d08d9e3ed00c066 Mon Sep 17 00:00:00 2001
From: Li Min <11663212+limin2021@users.noreply.github.com>
Date: Thu, 18 Nov 2021 16:21:03 +0800
Subject: [PATCH] fix bug to support dropout eval grad computing. (#37305)

* fix bug to support dropout eval grad computing.

* Remove useless code.
---
 paddle/fluid/operators/dropout_impl.cu.h   | 48 +++++++++++++---------
 paddle/fluid/operators/dropout_impl_util.h |  3 --
 paddle/fluid/operators/dropout_op.cu       |  8 ++--
 3 files changed, 31 insertions(+), 28 deletions(-)

diff --git a/paddle/fluid/operators/dropout_impl.cu.h b/paddle/fluid/operators/dropout_impl.cu.h
index 695d29b294a..bd4d690577a 100644
--- a/paddle/fluid/operators/dropout_impl.cu.h
+++ b/paddle/fluid/operators/dropout_impl.cu.h
@@ -244,34 +244,42 @@ void DropoutGradGPUKernelDriver(const platform::CUDADeviceContext& dev_ctx,
                                 const std::string dropout_implementation,
                                 float dropout_prob, const Tensor& grad_y,
                                 const Tensor& mask, int64_t size,
-                                Tensor* grad_x) {
-  auto M = EigenVector<uint8_t>::Flatten(mask);
+                                Tensor* grad_x, bool is_test = false) {
   auto dX = EigenVector<T>::Flatten(*grad_x);
   auto dY = EigenVector<T>::Flatten(grad_y);
 
   auto& place = *dev_ctx.eigen_device();
-  if (dropout_implementation == "upscale_in_train") {
-    if (dropout_prob == 1.0f) {
-      dX.device(place) = static_cast<T>(0) * dY;
+  if (is_test) {
+    if (dropout_implementation == "upscale_in_train") {
+      dX.device(place) = static_cast<T>(1) * dY;
     } else {
-      int vec_size = platform::GetVectorizedSize<T>(grad_y.data<T>());
-      if (vec_size == 4 && size % 4 == 0) {
-        auto factor = static_cast<T>(1.0f / (1.0f - dropout_prob));
-        auto stream = dev_ctx.stream();
-        platform::GpuLaunchConfig config =
-            platform::GetGpuLaunchConfig1D(dev_ctx, size);
-        DropoutGradCUDAKernel<
-            T, uint8_t,
-            4><<<config.block_per_grid, config.thread_per_block, 0, stream>>>(
-            grad_y.data<T>(), mask.data<uint8_t>(), factor, size,
-            grad_x->data<T>());
+      dX.device(place) = dY * static_cast<T>(1.0f - dropout_prob);
+    }
+  } else {
+    auto M = EigenVector<uint8_t>::Flatten(mask);
+    if (dropout_implementation == "upscale_in_train") {
+      if (dropout_prob == 1.0f) {
+        dX.device(place) = static_cast<T>(0) * dY;
       } else {
-        dX.device(place) =
-            dY * M.cast<T>() / static_cast<T>(1.0f - dropout_prob);
+        int vec_size = platform::GetVectorizedSize<T>(grad_y.data<T>());
+        if (vec_size == 4 && size % 4 == 0) {
+          auto factor = static_cast<T>(1.0f / (1.0f - dropout_prob));
+          auto stream = dev_ctx.stream();
+          platform::GpuLaunchConfig config =
+              platform::GetGpuLaunchConfig1D(dev_ctx, size);
+          DropoutGradCUDAKernel<
+              T, uint8_t,
+              4><<<config.block_per_grid, config.thread_per_block, 0, stream>>>(
+              grad_y.data<T>(), mask.data<uint8_t>(), factor, size,
+              grad_x->data<T>());
+        } else {
+          dX.device(place) =
+              dY * M.cast<T>() / static_cast<T>(1.0f - dropout_prob);
+        }
       }
+    } else {
+      dX.device(place) = dY * M.cast<T>();
     }
-  } else {
-    dX.device(place) = dY * M.cast<T>();
   }
 }
 
diff --git a/paddle/fluid/operators/dropout_impl_util.h b/paddle/fluid/operators/dropout_impl_util.h
index e11640d0706..f2038d12528 100644
--- a/paddle/fluid/operators/dropout_impl_util.h
+++ b/paddle/fluid/operators/dropout_impl_util.h
@@ -34,9 +34,6 @@ inline void GetSeedDataAndIncrement(const platform::CUDADeviceContext& dev_ctx,
     TensorCopySync(*seed, platform::CPUPlace(), &seed_cpu_tensor);
     *seed_data = static_cast<uint64_t>(seed_cpu_tensor.data<int>()[0]);
     *increment = offset;
-  } else if (seed && platform::is_cpu_place(seed->place())) {
-    *seed_data = *(seed->data<int>());
-    *increment = offset;
   } else if (gen_cuda->GetIsInitPy() && (!is_fix_seed)) {
     auto seed_offset = gen_cuda->IncrementOffset(offset);
     *seed_data = seed_offset.first;
diff --git a/paddle/fluid/operators/dropout_op.cu b/paddle/fluid/operators/dropout_op.cu
index 447184b9483..0d5ee41c5c3 100644
--- a/paddle/fluid/operators/dropout_op.cu
+++ b/paddle/fluid/operators/dropout_op.cu
@@ -58,10 +58,6 @@ template <typename DeviceContext, typename T>
 class GPUDropoutGradKernel : public framework::OpKernel<T> {
  public:
   void Compute(const framework::ExecutionContext& context) const override {
-    PADDLE_ENFORCE_EQ(!context.Attr<bool>("is_test"), true,
-                      platform::errors::PreconditionNotMet(
-                          "GradOp is only callable when is_test is false"));
-
     auto* grad_x = context.Output<Tensor>(framework::GradVarName("X"));
     auto* grad_y = context.Input<Tensor>(framework::GradVarName("Out"));
     auto* mask = context.Input<Tensor>("Mask");
@@ -71,10 +67,12 @@ class GPUDropoutGradKernel : public framework::OpKernel<T> {
         context.Attr<std::string>("dropout_implementation");
     float dropout_prob = context.Attr<float>("dropout_prob");
 
+    bool is_test = context.Attr<bool>("is_test");
+
     auto& dev_ctx =
         context.template device_context<platform::CUDADeviceContext>();
     DropoutGradGPUKernelDriver<T>(dev_ctx, dropout_implementation, dropout_prob,
-                                  *grad_y, *mask, size, grad_x);
+                                  *grad_y, *mask, size, grad_x, is_test);
   }
 };
 
-- 
GitLab