made changes according to feedback

src/tf/tf.cpp

@@ -36,7 +36,7 @@ struct tf_parser
 
     std::unordered_map<std::string, op_func> ops;
 
-    std::vector<size_t> parse_axes(attribute_map& attributes, const std::string& s) const
+    std::vector<size_t> parse_axes(const attribute_map& attributes, const std::string& s) const
     {
         auto attrs = attributes.at(s).list().i();
         std::vector<size_t> axes;
@@ -45,7 +45,7 @@ struct tf_parser
         {
             for(size_t& axis : axes)
             {
-                parse_axis(axis);
+                axis = parse_axis(axis);
             }
         }
         return axes;
@@ -57,27 +57,27 @@ struct tf_parser
         std::vector<T> new_data(prev_data.size());
         for(size_t i = 0; i < new_data.size(); i++)
         {
-            auto new_idx = i;
-            parse_axis(new_idx);
+            auto new_idx = parse_axis(i);
             new_data.at(new_idx) = prev_data.at(i);
         }
         prev_data = new_data;
     }
 
     template <class T>
-    void parse_axis(T& dim) const
+    T parse_axis(const T& dim) const
     {
         if(is_nhwc)
         {
             switch(dim)
             {
-            case 0: dim = 0; break;
-            case 1: dim = 2; break;
-            case 2: dim = 3; break;
-            case 3: dim = 1; break;
-            default: break;
+            case 0: return 0;
+            case 1: return 2;
+            case 2: return 3;
+            case 3: return 1;
+            default: return T{dim};
             }
         }
+        return T{dim};
     }
 
     std::vector<int64_t> get_axes(size_t num_axes) const
@@ -224,8 +224,7 @@ struct tf_parser
     {
         // get index for axis within args
         size_t axis_idx = attributes.at("N").i();
-        size_t axis     = args[axis_idx]->eval().at<int64_t>();
-        parse_axis(axis);
+        size_t axis     = parse_axis(args[axis_idx]->eval().at<int64_t>());
         op::concat op{axis};
         // return only first N arguments (assuming last index is the axis value)
         return prog.add_instruction(
@@ -263,7 +262,6 @@ struct tf_parser
             {
                 std::vector<size_t> padding;
                 copy(attributes.at("explicit_paddings").list().i(), std::back_inserter(padding));
-                reorder_data(padding);
                 if(padding.size() != 4)
                 {
                     MIGRAPHX_THROW("padding should have 4 values");
@@ -369,7 +367,7 @@ struct tf_parser
         op::reshape op;
         if(args.size() != 2)
             MIGRAPHX_THROW("reshape needs 2 arguments (input, new_shape)");
-        literal s = args[1]->get_literal();
+        auto s = args[1]->eval();
         s.visit([&](auto v) { copy(v, std::back_inserter(op.dims)); });
         return prog.add_instruction(op, args[0]);
     }
@@ -674,7 +672,12 @@ struct tf_parser
         {
             std::vector<int> data_int32 = get_data_vals(t.half_val(), shape_size);
             std::vector<uint16_t> data_uint16(data_int32.begin(), data_int32.end());
-            return literal{{shape::half_type, dims}, data_uint16};
+            std::vector<half> data_half;
+            std::transform(data_uint16.begin(),
+                           data_uint16.end(),
+                           std::back_inserter(data_half),
+                           [](uint16_t raw_val) { return *reinterpret_cast<half*>(&raw_val); });
+            return literal{{shape::half_type, dims}, data_half};
         }
         case tensorflow::DataType::DT_DOUBLE:
             return literal{{shape::double_type, dims}, get_data_vals(t.double_val(), shape_size)};