Removing stray code comments

llvm/projects/keras/frontend/approxhpvm_translator.py

@@ -80,7 +80,6 @@ class DFG:
         self.add_dfg_edge(layer.input[j].op.name, dfg_node)
 
     else:
-      #if self.hasSingleInput(layer):
       print (layer.input.name)        
       self.add_dfg_edge(layer.input.name, dfg_node)
 
@@ -183,21 +182,13 @@ class DFGNode:
         print ("padding = ", self.padding);
         
       if layer_type == "BatchNormalization":
-        #print (layer.__dir__())
         self.epsilon = layer.epsilon
         self.beta = layer.beta
         self.gamma = layer.gamma
         self.moving_mean = layer.moving_mean
         self.moving_variance = layer.moving_variance
         
-        #print("moving mean = ", self.moving_mean, "\n")
-        #print (self.moving_mean.__dir__())
-     
-        #print("moving variance = ", self.moving_variance, "\n")
-        #print (self.moving_variance.__dir__())
 
-        #print("gamma = ", self.gamma, "\n")
-        #print("beta = ", self.beta, "\n")
         
         
         
@@ -280,9 +271,8 @@ class TensorRtTranslator:
 
     print ("Input_type = ", cur_node.inputs[0].layer_type)
 
+    # NOTE: Assuming the 'inference' phase - hence skipping Dropout
     pred_layer_type = cur_node.inputs[0].layer_type
-    # FIXME: Assuming the 'inference' phase - hence skipping Dropout
-    #if pred_layer_type == "Flatten" or pred_layer_type == "Dropout":
     if self.isSkipLayer(pred_layer_type):
       cur_node = self.getPrevActiveLayer(cur_node)
 
@@ -726,9 +716,7 @@ class TensorRtTranslator:
     if test_data is not None and self.dfg.last_node is not None:
       last_node = self.dfg.last_node
       output_var = self.output_map[last_node.layer_name]
-      #accuracy_call = "\ncomputeAccuracy2(labels," + str(len(test_data)) + "," + output_var + "); \n"
-      #self.program_str += accuracy_call
-
+    
       
     destructors = "\nllvm_hpvm_cleanupTensorRt(); \n"
     self.program_str += destructors
@@ -804,8 +792,7 @@ class TensorRtTranslator:
   def endBatchLoop(self):
 
     end_loop_str = ""
-    #end_loop_str += "\nuint8_t* labels = readLabelsBatch(labels_path.c_str(),start,end); \n"
-    #end_loop_str += "\nuint32_t* labels = readLabelsBatch2(labels_path.c_str(),start,end); \n"
+    
     end_loop_str += "\nuint32_t* labels = readLabelsBatch3(labels_path.c_str(),start,end); \n"
 
     last_node = self.dfg.last_node

llvm/projects/keras/frontend/hpvm_dfg_translator.py

@@ -65,8 +65,9 @@ class HPVMTranslator:
       return "input"
 
     print ("Input_type = ", cur_node.inputs[0].layer_type)
+
     pred_layer_type = cur_node.inputs[0].layer_type
-    # FIXME: Assuming the 'inference' phase - hence skipping Dropout
+    # NOTE: Assuming the 'inference' phase - hence skipping Dropout
     #if pred_layer_type == "Flatten" or pred_layer_type == "Dropout":
     if self.isSkipLayer(pred_layer_type):
       cur_node = self.getPrevActiveLayer(cur_node)
@@ -228,7 +229,6 @@ class HPVMTranslator:
       hpvm_edge_str += "  __visc__bindIn(" + out_var_name + ", " + str(index2) + ", 1, 0); \n"
 
     elif input_var_name in self.hpvm_node_names:
-      #input_index = self.output_map[input_var_name]
       hpvm_edge_str += "  __visc__edge(" + input_var_name + ", " + out_var_name + ", 1, 0, 0, 0); \n"
       hpvm_edge_str += "  __visc__edge(" + input_var_name + ", " + out_var_name + ", 1, 1, 1, 0); \n"
 
@@ -241,7 +241,6 @@ class HPVMTranslator:
       hpvm_edge_str += "  __visc__bindIn(" + out_var_name + ", " + str(index2) + ", 3, 0); \n"
 
     elif input_var_name2 in self.hpvm_node_names:
-      #input_index = self.output_map[input_var_name2]
       hpvm_edge_str += "  __visc__edge(" + input_var_name2 + ", " + out_var_name + ", 1, 0, 2, 0); \n"
       hpvm_edge_str += "  __visc__edge(" + input_var_name2 + ", " + out_var_name + ", 1, 1, 3, 0); \n"
 
@@ -259,7 +258,7 @@ class HPVMTranslator:
     footer_str = self.genNodeFooter(2)
     inst_str += footer_str
 
-    #self.genHpvmNodeVar(out_var_name)
+    
     input_var_name = self.getSingleInputName(cur_node)
     weight_name = cur_node.layer_name + "_w"
     
@@ -497,11 +496,9 @@ class HPVMTranslator:
       self.genDenseNode(cur_node)      
 
     if nodeHasBias(cur_node):  
-    #if self.hasBiasAdd(cur_node):
       self.genBiasNode(cur_node)
 
     if nodeHasActivation(cur_node) and layer_type != "Activation":  
-    #if self.hasActivation(cur_node):
       self.genSubActivationNode(cur_node)     
       
     if layer_type == "Activation":
@@ -637,7 +634,6 @@ class HPVMTranslator:
     main_func_str += "hpvm_request_tensor(result, 0); \n\n"
     main_func_str += "__visc__cleanup(); \n "
    
-    #main_func_str += "computeAccuracy2(labels, " + str(len(test_data)) + ", result); \n"
     main_func_str += "computeAccuracy3(labels, result); \n"    
     main_func_str += "return 0; \n\n"
     main_func_str += "} \n"