commit (PaddlePaddle#44534)

Aurelius84 · Jul 29, 2022 · 363458b · 363458b
1 parent d2f3904
commit 363458b
Show file tree

Hide file tree

Showing 2 changed files with 86 additions and 18 deletions.
diff --git a/paddle/fluid/inference/tensorrt/convert/fc_op.cc b/paddle/fluid/inference/tensorrt/convert/fc_op.cc
@@ -333,23 +333,91 @@ class FcOpConverter : public OpConverter {
     if (!engine_->with_dynamic_shape()) {
       x_num_col_dims--;
     }
-    PADDLE_ENFORCE_GT(
-        x_dim.nbDims,
-        x_num_col_dims,
-        platform::errors::InvalidArgument(
-            "Params and input dims mismatch. Paddle-TRT FC "
-            "converter expects x_dim.nbDims > x_num_col_dims, but "
-            "x_dim.nbDims : %d, x_num_col_dims : %d.",
-            x_dim.nbDims,
-            x_num_col_dims));
-    // need reshape input before and after fc
-    auto* reshape_before_fc_layer =
-        reshape_before_fc(X, x_dim, x_num_col_dims, output_name);
-    auto* reshape_itensor = reshape_before_fc_layer->getOutput(0);
-    if (enable_int8 || support_int8) {
-      engine_->SetTensorDynamicRange(reshape_itensor, in_scale);
+    // If use tensorrt'oss, the x_dim and x_num_col_dims need change, and can
+    // not add Shuffle layer in ernie's multihead.
+    if (x_dim.nbDims == 4 && x_num_col_dims == 1) {
+      if (enable_int8 || support_int8) {
+        // add conv1x1 layer
+        nvinfer1::DimsHW nv_ksize(1, 1);
+        auto* fc_layer_int8 = TRT_ENGINE_ADD_LAYER(engine_,
+                                                   Convolution,
+                                                   *X,
+                                                   n_output,
+                                                   nv_ksize,
+                                                   weight.get(),
+                                                   bias.get());
+        if (activation_type == "relu") {
+          fc_layer_int8->setName(
+              ("ernie_fc_op_int8: Convolution (Output: " + output_name + ")")
+                  .c_str());
+          PADDLE_ENFORCE_EQ(
+              op_desc.HasAttr("out_threshold"),
+              true,
+              platform::errors::InvalidArgument(
+                  "must have out threshold in fc layers in int8 mode"));
+          float out_scale = 0;
+          if (enable_int8) {
+            out_scale =
+                PADDLE_GET_CONST(float, op_desc.GetAttr("out_threshold"));
+          } else {
+            out_scale = PADDLE_GET_CONST(float, op_desc.GetAttr("Out"));
+          }
+          engine_->SetTensorDynamicRange(fc_layer_int8->getOutput(0),
+                                         out_scale);
+          nvinfer1::IActivationLayer* relu_layer_int8 =
+              TRT_ENGINE_ADD_LAYER(engine_,
+                                   Activation,
+                                   *(fc_layer_int8->getOutput(0)),
+                                   nvinfer1::ActivationType::kRELU);
+          RreplenishLayerAndOutput(relu_layer_int8,
+                                   "relu_after_ernie_fc_int8",
+                                   {output_name},
+                                   test_mode);
+        } else {
+          RreplenishLayerAndOutput(fc_layer_int8,
+                                   "ernie_fc_op_int8: Convolution",
+                                   {output_name},
+                                   test_mode);
+        }
+      } else {
+        // add fc layer
+        auto* fc_layer_float = TRT_ENGINE_ADD_LAYER(
+            engine_, FullyConnected, *X, n_output, weight.get(), bias.get());
+        if (activation_type == "relu") {
+          fc_layer_float->setName(
+              ("ernie_fc_op_float: (Output: " + output_name + ")").c_str());
+          nvinfer1::IActivationLayer* relu_layer_float =
+              TRT_ENGINE_ADD_LAYER(engine_,
+                                   Activation,
+                                   *(fc_layer_float->getOutput(0)),
+                                   nvinfer1::ActivationType::kRELU);
+          RreplenishLayerAndOutput(relu_layer_float,
+                                   "relu_after_ernie_fc_float",
+                                   {output_name},
+                                   test_mode);
+        } else {
+          RreplenishLayerAndOutput(
+              fc_layer_float, "ernie_fc_op_float", {output_name}, test_mode);
+        }
+      }
+    } else {  // need reshape input before and after fc
+      PADDLE_ENFORCE_GT(
+          x_dim.nbDims,
+          x_num_col_dims,
+          platform::errors::InvalidArgument(
+              "Params and input dims mismatch. Paddle-TRT FC "
+              "converter expects x_dim.nbDims > x_num_col_dims, but "
+              "x_dim.nbDims : %d, x_num_col_dims : %d.",
+              x_dim.nbDims,
+              x_num_col_dims));
+      auto* reshape_before_fc_layer =
+          reshape_before_fc(X, x_dim, x_num_col_dims, output_name);
+      auto* reshape_itensor = reshape_before_fc_layer->getOutput(0);
+      if (enable_int8 || support_int8) {
+        engine_->SetTensorDynamicRange(reshape_itensor, in_scale);
+      }
+      regist_fc(reshape_itensor, n_output, weight, bias);
     }
-    regist_fc(reshape_itensor, n_output, weight, bias);
   }
 };
 

diff --git a/python/paddle/fluid/tests/unittests/ir/inference/test_trt_convert_fc.py b/python/paddle/fluid/tests/unittests/ir/inference/test_trt_convert_fc.py
@@ -249,9 +249,9 @@ def test(self):
 
 # this is the special case when x_dim.nbDims == 4 && x_num_col_dims == 1
 class TrtConvertFcTest3(TrtLayerAutoScanTest):
-
+    # this case will invoke a bug in fc_op.cc, so return False
     def is_program_valid(self, program_config: ProgramConfig) -> bool:
-        return True
+        return False
 
     def sample_program_configs(self):
         self.trt_param.workspace_size = 1073741824