From 2f3ad5abd0c53e374889253dd3c337fd4a83046d Mon Sep 17 00:00:00 2001
From: feng_shuai <fengshuai03@baidu.com>
Date: Mon, 24 Oct 2022 15:54:48 +0800
Subject: [PATCH] optimize: vit static shape (#47280)

---
 .../tensorrt/convert/multihead_matmul_op.cc   | 19 +++++++++----
 .../test_trt_convert_multihead_matmul.py      | 28 ++++++++++++++++++-
 2 files changed, 41 insertions(+), 6 deletions(-)

diff --git a/paddle/fluid/inference/tensorrt/convert/multihead_matmul_op.cc b/paddle/fluid/inference/tensorrt/convert/multihead_matmul_op.cc
index f5af2a967c..f997c8bd1f 100644
--- a/paddle/fluid/inference/tensorrt/convert/multihead_matmul_op.cc
+++ b/paddle/fluid/inference/tensorrt/convert/multihead_matmul_op.cc
@@ -460,9 +460,10 @@ class MultiheadMatMulOpConverter : public OpConverter {
           plugin_inputs.emplace_back(mask_tensor);
           // input_2 for plugin
           std::vector<int> pos_id = {0};
-          int max_batch = 500;
+          int max_batch = 512;
+          int length = (input_dims.d[1] == -1) ? 1 : input_dims.d[1];
           for (int i = 1; i < max_batch; i++) {
-            pos_id.push_back(i);
+            pos_id.push_back(i * length);
           }
           nvinfer1::ITensor* fake_pos_id_tensor = Add1DConstantLayer(pos_id);
           nvinfer1::ITensor* length_tensor =
@@ -497,18 +498,26 @@ class MultiheadMatMulOpConverter : public OpConverter {
           stride.d[0] = 1;
           size.d[0] = 1;
 
+          nvinfer1::ITensor* pos_id_tensor = (input_dims.d[1] == -1)
+                                                 ? pos_id_layer->getOutput(0)
+                                                 : fake_pos_id_tensor;
+
           auto* slice_pos_layer = TRT_ENGINE_ADD_LAYER(
-              engine_, Slice, *pos_id_layer->getOutput(0), start, size, stride);
+              engine_, Slice, *pos_id_tensor, start, size, stride);
           slice_pos_layer->setInput(2, *size_layer->getOutput(0));
           plugin_inputs.emplace_back(slice_pos_layer->getOutput(0));
 
           // input_3 for plugin
-          std::vector<int> data(500, 1);
+          int max_length = (input_dims.d[1] == -1) ? 512 : input_dims.d[1];
+          std::vector<int> data(max_length, 1);
           nvinfer1::ITensor* fake_max_seqlen_tensor = Add1DConstantLayer(data);
           auto* slice_max_layer = TRT_ENGINE_ADD_LAYER(
               engine_, Slice, *fake_max_seqlen_tensor, start, size, stride);
           slice_max_layer->setInput(2, *length_tensor);
-          plugin_inputs.emplace_back(slice_max_layer->getOutput(0));
+          nvinfer1::ITensor* max_seqlen_tensor =
+              (input_dims.d[1] == -1) ? slice_max_layer->getOutput(0)
+                                      : fake_max_seqlen_tensor;
+          plugin_inputs.emplace_back(max_seqlen_tensor);
           // plugin_layer
           auto plugin_layer = engine_->network()->addPluginV2(
               plugin_inputs.data(), plugin_inputs.size(), *plugin);
diff --git a/python/paddle/fluid/tests/unittests/ir/inference/test_trt_convert_multihead_matmul.py b/python/paddle/fluid/tests/unittests/ir/inference/test_trt_convert_multihead_matmul.py
index db02ce1e0f..6889f88fa4 100644
--- a/python/paddle/fluid/tests/unittests/ir/inference/test_trt_convert_multihead_matmul.py
+++ b/python/paddle/fluid/tests/unittests/ir/inference/test_trt_convert_multihead_matmul.py
@@ -808,7 +808,7 @@ class TrtConvertVitToMultiHeadMatmulTest(TrtLayerAutoScanTest):
 
         for batch in [2, 4]:
             self.batch = batch
-            for length in [64, 384]:
+            for length in [197]:
                 self.length = length
                 ops_config = [
                     {
@@ -1006,6 +1006,17 @@ class TrtConvertVitToMultiHeadMatmulTest(TrtLayerAutoScanTest):
                 "input_data1": [1, 197, 768],
             }
 
+        def generate_static_shape(attrs):
+            self.dynamic_shape.min_input_shape = {
+                "input_data1": [1, 197, 768],
+            }
+            self.dynamic_shape.max_input_shape = {
+                "input_data1": [16, 197, 768],
+            }
+            self.dynamic_shape.opt_input_shape = {
+                "input_data1": [1, 197, 768],
+            }
+
         def clear_dynamic_shape():
             self.dynamic_shape.max_input_shape = {}
             self.dynamic_shape.min_input_shape = {}
@@ -1035,6 +1046,21 @@ class TrtConvertVitToMultiHeadMatmulTest(TrtLayerAutoScanTest):
             1e-5,
         )
 
+        # for static_shape
+        clear_dynamic_shape()
+        generate_static_shape(attrs)
+        self.trt_param.workspace_size = 2013265920
+        self.trt_param.precision = paddle_infer.PrecisionType.Half
+        yield self.create_inference_config(), generate_trt_nodes_num(), (
+            1e-3,
+            1e-3,
+        )
+        self.trt_param.precision = paddle_infer.PrecisionType.Float32
+        yield self.create_inference_config(), generate_trt_nodes_num(), (
+            1e-5,
+            1e-5,
+        )
+
     def add_skip_trt_case(self):
         pass
 
-- 
GitLab