Correct MultipleQuantizeSquash (#40717)

* Correct MultipleQuantizeSquash

* Correct logging

paddle/fluid/framework/ir/mkldnn/cpu_quantize_pass.cc

@@ -39,12 +39,13 @@ void UnlinkNodes(ir::Node* a, ir::Node* b) {
                   b->inputs.end());
 }
 
-void LogCannotQuantizeOp(Node* op, const char* details = nullptr) {
+void MarkAndLogCannotQuantizeOp(Node* op, const char* details = nullptr) {
   std::stringstream msg_ss;
   msg_ss << "Cannot quantize operator " << op->Name()
          << " (type: " << op->Op()->Type() << ", id: " << op->id() << ").";
   if (details) msg_ss << " " << details;
-  PrettyLogDetail(msg_ss.str().c_str());
+  VLOG(2) << msg_ss.str().c_str();
+  op->Op()->SetAttr("mkldnn_data_type", std::string("float32"));
 }
 
 void LogScaleIsMissingForVarName(const std::string& name) {
@@ -56,12 +57,19 @@ void LogScaleIsMissingForVarNode(Node* node) {
 }
 
 void LogQuantizationDisabled(Node* op) {
-  std::stringstream msg_ss;
-  VLOG(4) << "Qantization skipped for operator " << op->Name()
+  VLOG(2) << "Quantization skipped for operator " << op->Name()
           << " (type: " << op->Op()->Type() << ", id: " << op->id()
           << "). Attribute mkldnn_data_type != \"int8\".";
 }
 
+void LogQuantizedOpsCounter(const std::string& type, const int counter,
+                            const char* details = nullptr) {
+  std::stringstream msg_ss;
+  msg_ss << "---    quantized " << counter << " " << type << " ops";
+  if (details) msg_ss << " " << details;
+  PrettyLogDetail(msg_ss.str().c_str());
+}
+
 }  // namespace
 
 enum { U8_MAX = 255, S8_MAX = 127 };
@@ -307,7 +315,8 @@ void CPUQuantizePass::QuantizeConv(Graph* graph,
 
     auto has_output_scale = AreScalesPresentForNodes({conv_output});
     if (with_residual_data && !has_output_scale) {
-      LogCannotQuantizeOp(conv_op,
+      MarkAndLogCannotQuantizeOp(
+          conv_op,
           "Conv op with ResidualData input cannot be quantized "
           "without output scale.");
       return;
@@ -318,7 +327,8 @@ void CPUQuantizePass::QuantizeConv(Graph* graph,
                                 conv_pattern);
       if (!AreScalesPresentForNodes(
               {conv_input, conv_filter, conv_residual_data})) {
-        LogCannotQuantizeOp(conv_op, "No scale available for the operator");
+        MarkAndLogCannotQuantizeOp(conv_op,
+                                   "No scale available for the operator");
         return;
       }
 
@@ -330,7 +340,8 @@ void CPUQuantizePass::QuantizeConv(Graph* graph,
                     residual_scale, is_residual_unsigned, "Scale_in_eltwise");
     } else {
       if (!AreScalesPresentForNodes({conv_input, conv_filter})) {
-        LogCannotQuantizeOp(conv_op, "No scale available for the operator");
+        MarkAndLogCannotQuantizeOp(conv_op,
+                                   "No scale available for the operator");
         return;
       }
     }
@@ -377,10 +388,9 @@ void CPUQuantizePass::QuantizeConv(Graph* graph,
   gpd(graph, handler);
   AddStatis(quantize_conv_count);
 
-  std::stringstream msg_ss;
-  msg_ss << "---    quantized " << quantize_conv_count << " conv2d ops";
-  if (with_residual_data) msg_ss << " with residual connection";
-  PrettyLogDetail(msg_ss.str().c_str());
+  LogQuantizedOpsCounter(
+      "conv2d", quantize_conv_count,
+      ((with_residual_data) ? "with residual connection" : ""));
 }
 
 void CPUQuantizePass::QuantizeFc(Graph* graph) const {
@@ -405,7 +415,7 @@ void CPUQuantizePass::QuantizeFc(Graph* graph) const {
       return;
     }
     if (!fc->Op()->GetAttrIfExists<bool>("use_mkldnn")) {
-      LogCannotQuantizeOp(fc, "use_mkldnn attribute set to false");
+      MarkAndLogCannotQuantizeOp(fc, "use_mkldnn attribute set to false");
       return;
     }
 
@@ -414,7 +424,7 @@ void CPUQuantizePass::QuantizeFc(Graph* graph) const {
     GET_IR_NODE_FROM_SUBGRAPH(output, output, fc_pattern);
 
     if (!AreScalesPresentForNodes({input, weights})) {
-      LogCannotQuantizeOp(fc, "No scale available for the operator");
+      MarkAndLogCannotQuantizeOp(fc, "No scale available for the operator");
       return;
     }
 
@@ -448,10 +458,7 @@ void CPUQuantizePass::QuantizeFc(Graph* graph) const {
 
   gpd(graph, handler);
   AddStatis(quantize_fc_count);
-
-  std::stringstream msg_ss;
-  msg_ss << "---    quantized " << quantize_fc_count << " fc ops";
-  PrettyLogDetail(msg_ss.str().c_str());
+  LogQuantizedOpsCounter("fc", quantize_fc_count);
 }
 
 void CPUQuantizePass::QuantizePool(Graph* graph) const {
@@ -476,7 +483,8 @@ void CPUQuantizePass::QuantizePool(Graph* graph) const {
     GET_IR_NODE_FROM_SUBGRAPH(pool_output, pool_output, pool_pattern);
 
     if (!AreScalesPresentForNodes({pool_input, pool_output})) {
-      LogCannotQuantizeOp(pool_op, "No scale available for the operator");
+      MarkAndLogCannotQuantizeOp(pool_op,
+                                 "No scale available for the operator");
       return;
     }
 
@@ -494,8 +502,7 @@ void CPUQuantizePass::QuantizePool(Graph* graph) const {
 
   gpd(graph, handler);
   AddStatis(quantize_pool_count);
-
-  PrettyLogDetail("---    quantized %d pool2d ops", quantize_pool_count);
+  LogQuantizedOpsCounter("pool2d", quantize_pool_count);
 }
 
 void CPUQuantizePass::QuantizeConcat(Graph* graph) const {
@@ -519,7 +526,8 @@ void CPUQuantizePass::QuantizeConcat(Graph* graph) const {
     GET_IR_NODE_FROM_SUBGRAPH(concat_out, concat_out, concat_pattern);
 
     if (!AreScalesPresentForNodes({concat_out})) {
-      LogCannotQuantizeOp(concat_op, "No scale available for the operator");
+      MarkAndLogCannotQuantizeOp(concat_op,
+                                 "No scale available for the operator");
       return;
     }
 
@@ -539,8 +547,7 @@ void CPUQuantizePass::QuantizeConcat(Graph* graph) const {
 
   gpd(graph, handler);
   AddStatis(quantize_concat_count);
-
-  PrettyLogDetail("---    quantized %d concat ops", quantize_concat_count);
+  LogQuantizedOpsCounter("concat", quantize_concat_count);
 }
 
 void CPUQuantizePass::QuantizePriorBox(Graph* graph) const {
@@ -565,7 +572,8 @@ void CPUQuantizePass::QuantizePriorBox(Graph* graph) const {
                               prior_box_pattern);
 
     if (!AreScalesPresentForNodes({prior_box_input})) {
-      LogCannotQuantizeOp(prior_box_op, "No scale available for the operator");
+      MarkAndLogCannotQuantizeOp(prior_box_op,
+                                 "No scale available for the operator");
       return;
     }
 
@@ -580,9 +588,7 @@ void CPUQuantizePass::QuantizePriorBox(Graph* graph) const {
 
   gpd(graph, handler);
   AddStatis(quantize_prior_box_count);
-
-  PrettyLogDetail("---    quantized %d prior_box ops",
-                  quantize_prior_box_count);
+  LogQuantizedOpsCounter("prior_box", quantize_prior_box_count);
 }
 
 void CPUQuantizePass::QuantizeTranspose(Graph* graph) const {
@@ -608,13 +614,14 @@ void CPUQuantizePass::QuantizeTranspose(Graph* graph) const {
 
     // skip if prev op and next op is not quantized
     if (!(IsOpDequantized(prev_op)) && !(IsOpQuantized(transpose_out))) {
-      LogCannotQuantizeOp(transpose_op,
+      MarkAndLogCannotQuantizeOp(transpose_op,
                                  "No other quantizable operators nearby");
       return;
     }
 
     if (!AreScalesPresentForNodes({transpose_in, transpose_out})) {
-      LogCannotQuantizeOp(transpose_op, "No scale available for the operator");
+      MarkAndLogCannotQuantizeOp(transpose_op,
+                                 "No scale available for the operator");
       return;
     }
 
@@ -634,9 +641,7 @@ void CPUQuantizePass::QuantizeTranspose(Graph* graph) const {
 
   gpd(graph, handler);
   AddStatis(quantize_transpose_count);
-
-  PrettyLogDetail("---    quantized %d transpose ops",
-                  quantize_transpose_count);
+  LogQuantizedOpsCounter("transpose2", quantize_transpose_count);
 }
 
 void CPUQuantizePass::QuantizeReshape(Graph* graph) const {
@@ -662,12 +667,14 @@ void CPUQuantizePass::QuantizeReshape(Graph* graph) const {
 
     // skip if prev op is not quantized
     if (!(IsOpDequantized(prev_op)) && !(IsOpQuantized(reshape_out))) {
-      LogCannotQuantizeOp(reshape_op, "No other quantizable operators nearby");
+      MarkAndLogCannotQuantizeOp(reshape_op,
+                                 "No other quantizable operators nearby");
       return;
     }
 
     if (!AreScalesPresentForNodes({reshape_in, reshape_out})) {
-      LogCannotQuantizeOp(reshape_op, "No scale available for the operator");
+      MarkAndLogCannotQuantizeOp(reshape_op,
+                                 "No scale available for the operator");
       return;
     }
 
@@ -686,8 +693,7 @@ void CPUQuantizePass::QuantizeReshape(Graph* graph) const {
 
   gpd(graph, handler);
   AddStatis(quantize_reshape_count);
-
-  PrettyLogDetail("---    quantized %d reshape ops", quantize_reshape_count);
+  LogQuantizedOpsCounter("reshape2", quantize_reshape_count);
 }
 
 void CPUQuantizePass::QuantizeSlice(Graph* graph) const {
@@ -713,12 +719,14 @@ void CPUQuantizePass::QuantizeSlice(Graph* graph) const {
 
     // skip if prev op and next op is not quantized
     if (!IsOpDequantized(prev_op) && !IsOpQuantized(slice_out)) {
-      LogCannotQuantizeOp(slice_op, "No other quantizable operators nearby");
+      MarkAndLogCannotQuantizeOp(slice_op,
+                                 "No other quantizable operators nearby");
       return;
     }
 
     if (!AreScalesPresentForNodes({slice_out})) {
-      LogCannotQuantizeOp(slice_op, "No scale available for the operator");
+      MarkAndLogCannotQuantizeOp(slice_op,
+                                 "No scale available for the operator");
       return;
     }
 
@@ -737,8 +745,7 @@ void CPUQuantizePass::QuantizeSlice(Graph* graph) const {
 
   gpd(graph, handler);
   AddStatis(quantize_slice_count);
-
-  PrettyLogDetail("---    quantized %d slice ops", quantize_slice_count);
+  LogQuantizedOpsCounter("slice", quantize_slice_count);
 }
 
 void CPUQuantizePass::QuantizeMatmul(Graph* graph) const {
@@ -763,7 +770,8 @@ void CPUQuantizePass::QuantizeMatmul(Graph* graph) const {
 
     // skip if prev ops are not quantized
     if (!IsOpDequantized(prev_op_x) || !IsOpDequantized(prev_op_y)) {
-      LogCannotQuantizeOp(matmul_op, "No other quantizable operators nearby");
+      MarkAndLogCannotQuantizeOp(matmul_op,
+                                 "No other quantizable operators nearby");
       return;
     }
     GET_IR_NODE_FROM_SUBGRAPH(matmul_in_x, matmul_in_x, matmul_pattern);
@@ -771,7 +779,8 @@ void CPUQuantizePass::QuantizeMatmul(Graph* graph) const {
     GET_IR_NODE_FROM_SUBGRAPH(matmul_out, matmul_out, matmul_pattern);
 
     if (!AreScalesPresentForNodes({matmul_in_x, matmul_in_y})) {
-      LogCannotQuantizeOp(matmul_op, "No scale available for the operator");
+      MarkAndLogCannotQuantizeOp(matmul_op,
+                                 "No scale available for the operator");
       return;
     }
 
@@ -803,8 +812,7 @@ void CPUQuantizePass::QuantizeMatmul(Graph* graph) const {
   };
   gpd(graph, handler);
   AddStatis(quantize_matmul_count);
-
-  PrettyLogDetail("---    quantized %d matmul ops", quantize_matmul_count);
+  LogQuantizedOpsCounter("matmul", quantize_matmul_count);
 }
 
 void CPUQuantizePass::QuantizeElementwise(
@@ -840,7 +848,7 @@ void CPUQuantizePass::QuantizeElementwise(
 
     if (!AreScalesPresentForNodes(
             {elementwise_x, elementwise_y, elementwise_out})) {
-      LogCannotQuantizeOp(elementwise_op,
+      MarkAndLogCannotQuantizeOp(elementwise_op,
                                  "No scale available for the operator");
       return;
     }
@@ -851,8 +859,8 @@ void CPUQuantizePass::QuantizeElementwise(
 
     // TODO(sfraczek): add support for different signness
     if (is_x_unsigned != is_y_unsigned) {
-      LogCannotQuantizeOp(elementwise_op,
-                          "Elementwise inputs must be of the same type.");
+      MarkAndLogCannotQuantizeOp(
+          elementwise_op, "Elementwise inputs must be of the same type.");
       return;
     }
 
@@ -872,9 +880,7 @@ void CPUQuantizePass::QuantizeElementwise(
   };
   gpd(graph, handler);
   AddStatis(quantize_elementwise_count);
-
-  PrettyLogDetail("---    quantized %d %s ops", quantize_elementwise_count,
-                  elementwise_type);
+  LogQuantizedOpsCounter(elementwise_type, quantize_elementwise_count);
 }
 
 void CPUQuantizePass::QuantizeFusionGru(Graph* graph) const {
@@ -900,7 +906,7 @@ void CPUQuantizePass::QuantizeFusionGru(Graph* graph) const {
     GET_IR_NODE_FROM_SUBGRAPH(out, out, pattern);
 
     if (!AreScalesPresentForNodes({x, weight_x})) {
-      LogCannotQuantizeOp(op, "No scale available for the operator");
+      MarkAndLogCannotQuantizeOp(op, "No scale available for the operator");
       return;
     }
 
@@ -929,8 +935,7 @@ void CPUQuantizePass::QuantizeFusionGru(Graph* graph) const {
   };
   gpd(graph, handler);
   AddStatis(quantize_count);
-
-  PrettyLogDetail("---    quantized %d fusion_gru ops", quantize_count);
+  LogQuantizedOpsCounter("fusion_gru", quantize_count);
 }
 
 void CPUQuantizePass::QuantizeMultiGru(Graph* graph) const {
@@ -957,7 +962,7 @@ void CPUQuantizePass::QuantizeMultiGru(Graph* graph) const {
     auto wx_names = gru->Op()->Input("WeightX");
     if (!AreScalesPresentForNodes({x}) ||
         !AreScalesPresentForVarNames(wx_names)) {
-      LogCannotQuantizeOp(gru, "No scale available for the operator");
+      MarkAndLogCannotQuantizeOp(gru, "No scale available for the operator");
       return;
     }
 
@@ -1007,8 +1012,7 @@ void CPUQuantizePass::QuantizeMultiGru(Graph* graph) const {
   };
   gpd(graph, handler);
   AddStatis(quantize_count);
-
-  PrettyLogDetail("---    quantized %d multi_gru ops", quantize_count);
+  LogQuantizedOpsCounter("multi_gru", quantize_count);
 }
 
 void CPUQuantizePass::QuantizeFusionLSTM(Graph* graph) const {
@@ -1036,7 +1040,7 @@ void CPUQuantizePass::QuantizeFusionLSTM(Graph* graph) const {
 
     // Starting from here there maybe issues
     if (!AreScalesPresentForNodes({x, weight_x})) {
-      LogCannotQuantizeOp(op, "No scale available for the operator");
+      MarkAndLogCannotQuantizeOp(op, "No scale available for the operator");
       return;
     }
 
@@ -1065,8 +1069,7 @@ void CPUQuantizePass::QuantizeFusionLSTM(Graph* graph) const {
   };
   gpd(graph, handler);
   AddStatis(quantize_count);
-
-  PrettyLogDetail("---    quantized %d fusion_lstm ops", quantize_count);
+  LogQuantizedOpsCounter("fusion_lstm", quantize_count);
 }
 
 void CPUQuantizePass::QuantizeNearestInterp(Graph* graph) const {
@@ -1095,13 +1098,13 @@ void CPUQuantizePass::QuantizeNearestInterp(Graph* graph) const {
 
     // skip if prev op and next op is not quantized
     if (!(IsOpDequantized(prev_op)) && !(IsOpQuantized(nearest_interp_out))) {
-      LogCannotQuantizeOp(nearest_interp_op,
+      MarkAndLogCannotQuantizeOp(nearest_interp_op,
                                  "No other quantizable operators nearby");
       return;
     }
 
     if (!AreScalesPresentForNodes({nearest_interp_in, nearest_interp_out})) {
-      LogCannotQuantizeOp(nearest_interp_op,
+      MarkAndLogCannotQuantizeOp(nearest_interp_op,
                                  "No scale available for the operator");
       return;
     }
@@ -1123,9 +1126,7 @@ void CPUQuantizePass::QuantizeNearestInterp(Graph* graph) const {
 
   gpd(graph, handler);
   AddStatis(quantize_nearest_interp_count);
-
-  PrettyLogDetail("---    quantized %d nearest_interp ops",
-                  quantize_nearest_interp_count);
+  LogQuantizedOpsCounter("nearest_interp", quantize_nearest_interp_count);
 }
 
 void CPUQuantizePass::ApplyImpl(ir::Graph* graph) const {

paddle/fluid/framework/ir/mkldnn/cpu_quantize_squash_pass.cc

@@ -434,9 +434,17 @@ void CPUQuantizeSquashPass::MultipleQuantizeSquash(Graph* graph) const {
             platform::errors::NotFound("Operator after quantize operator(%s) "
                                        "should has quantize output as input.",
                                        quant_out->Name()));
-        last_op->Op()->SetInput(
-            last_op_input_name,
-            std::vector<std::string>({first_quant_out->Name()}));
+
+        // update the next operator input,
+        // by replacing quant_out with first_quant_out
+        auto last_op_names = last_op->Op()->Input(last_op_input_name);
+        last_op_names.erase(std::remove(last_op_names.begin(),
+                                        last_op_names.end(), quant_out->Name()),
+                            last_op_names.end());
+        last_op_names.push_back(first_quant_out->Name());
+        last_op->Op()->SetInput(last_op_input_name,
+                                std::vector<std::string>(last_op_names));
+
         IR_NODE_LINK_TO(first_quant_out, last_op);
         GraphSafeRemoveNodes(graph, {quant_op, quant_out});
         removed_quantize++;