From a0a7a7e7025e2257a4b264c7ec607bfd45dc5586 Mon Sep 17 00:00:00 2001
From: Zheng-Bicheng <zheng_bicheng@outlook.com>
Date: Fri, 20 Sep 2024 10:43:00 +0800
Subject: [PATCH 1/2] update

---
 paddle2onnx/mapper/quantize/rknn_quantize_processor.cc | 2 --
 1 file changed, 2 deletions(-)

diff --git a/paddle2onnx/mapper/quantize/rknn_quantize_processor.cc b/paddle2onnx/mapper/quantize/rknn_quantize_processor.cc
index 1bc019ba2..a54dd58b9 100644
--- a/paddle2onnx/mapper/quantize/rknn_quantize_processor.cc
+++ b/paddle2onnx/mapper/quantize/rknn_quantize_processor.cc
@@ -169,13 +169,11 @@ void RKNNQuantizeProcessor::PerchannelToPerlayer() {
 
     auto next_nodes = name2node_dict_[node->output(0)];
     if (next_nodes.size() > 1 || IsGraphOutput(node->output(0))) {
-      P2OLogger() << "Type1" << std::endl;
       continue;
     }
 
     auto add_node = next_nodes[0];
     if (add_node->op_type() != "Add" || IsGraphOutput(add_node->output(0))) {
-      P2OLogger() << "Type2" << std::endl;
       continue;
     }
 

From 1d27f1884d413f060331b1656d4f29293940dc01 Mon Sep 17 00:00:00 2001
From: Zheng-Bicheng <zheng_bicheng@outlook.com>
Date: Fri, 20 Sep 2024 11:10:13 +0800
Subject: [PATCH 2/2] update

---
 .../quantize/rknn_quantize_processor.cc       | 30 ++++++++++++-------
 1 file changed, 20 insertions(+), 10 deletions(-)

diff --git a/paddle2onnx/mapper/quantize/rknn_quantize_processor.cc b/paddle2onnx/mapper/quantize/rknn_quantize_processor.cc
index a54dd58b9..af7af5f68 100644
--- a/paddle2onnx/mapper/quantize/rknn_quantize_processor.cc
+++ b/paddle2onnx/mapper/quantize/rknn_quantize_processor.cc
@@ -90,23 +90,33 @@ void RKNNQuantizeProcessor::AddQDQ() {
         if (helper_->quantize_info.find(name) != helper_->quantize_info.end()) {
           continue;
         }
-        std::vector<float> matmul_weight;
-        if (!GetTensorByName(name, &matmul_weight)) {
+        std::vector<float> weight_data;
+        if (!GetTensorByName(name, &weight_data)) {
           continue;
         }
-        std::vector<int64_t> matmul_weight_shape;
-        if (!GetTensorShape(name, &matmul_weight_shape)) {
+        std::vector<int64_t> weight_shape;
+        if (!GetTensorShape(name, &weight_shape)) {
           continue;
         }
+
         int64_t quantize_axis = 1;
         std::vector<float> scale;
         std::vector<int64_t> zeros;
-        GetChannelWiseQuantizeInfo(matmul_weight, matmul_weight_shape,
-                                   quantize_axis, &scale, &zeros);
-        auto scale_node =
-            helper_->Constant(ONNX_NAMESPACE::TensorProto::FLOAT, scale);
-        auto zero_node =
-            helper_->Constant(ONNX_NAMESPACE::TensorProto::INT8, zeros);
+        std::string scale_node, zero_node;
+        if (weight_shape.size() <= 1) {
+          GetTensorWiseQuantizeInfo(weight_data, &scale, &zeros);
+          scale_node = helper_->Constant({}, ONNX_NAMESPACE::TensorProto::FLOAT,
+                                         scale[0]);
+          zero_node = helper_->Constant({}, ONNX_NAMESPACE::TensorProto::INT8,
+                                        zeros[0]);
+        } else {
+          GetChannelWiseQuantizeInfo(weight_data, weight_shape, quantize_axis,
+                                     &scale, &zeros);
+          scale_node =
+              helper_->Constant(ONNX_NAMESPACE::TensorProto::FLOAT, scale);
+          zero_node =
+              helper_->Constant(ONNX_NAMESPACE::TensorProto::INT8, zeros);
+        }
         QuantizeInfo matmul_weight_quantize_info(scale, zeros, scale_node,
                                                  zero_node, quantize_axis);
         helper_->quantize_info[name] = matmul_weight_quantize_info;