Merge branch 'develop' into dot-add

test/onnx/onnx_test.cpp

@@ -3378,13 +3378,127 @@ TEST_CASE(nms_test)
     auto st = mm->add_parameter("score_threshold", sst);
 
     auto ret = mm->add_instruction(
-        migraphx::make_op("nonmaxsuppression", {{"center_point_box", 1}}), b, s, mo, iou, st);
+        migraphx::make_op("nonmaxsuppression", {{"center_point_box", true}}), b, s, mo, iou, st);
     mm->add_return({ret});
 
     auto prog = migraphx::parse_onnx("nms_test.onnx");
     EXPECT(p == prog);
 }
 
+TEST_CASE(nms_dynamic_batch_test)
+{
+    migraphx::program p;
+    auto* mm = p.get_main_module();
+    migraphx::shape sb{migraphx::shape::float_type, {{1, 10, 0}, {6, 6, 0}, {4, 4, 0}}};
+    auto b = mm->add_parameter("boxes", sb);
+    migraphx::shape ss{migraphx::shape::float_type, {{1, 10, 0}, {1, 1, 0}, {6, 6, 0}}};
+    auto s = mm->add_parameter("scores", ss);
+    migraphx::shape smo{migraphx::shape::int64_type, {1}};
+    auto mo = mm->add_parameter("max_output_boxes_per_class", smo);
+    migraphx::shape siou{migraphx::shape::float_type, {1}};
+    auto iou = mm->add_parameter("iou_threshold", siou);
+    migraphx::shape sst{migraphx::shape::float_type, {1}};
+    auto st  = mm->add_parameter("score_threshold", sst);
+    auto ret = mm->add_instruction(
+        migraphx::make_op("nonmaxsuppression",
+                          {{"center_point_box", true}, {"use_dyn_output", true}}),
+        b,
+        s,
+        mo,
+        iou,
+        st);
+    mm->add_return({ret});
+
+    migraphx::onnx_options options;
+    options.default_dyn_dim_value = {1, 10, 0};
+    options.use_dyn_output        = true;
+
+    auto prog = migraphx::parse_onnx("nms_dynamic_batch_test.onnx", options);
+    EXPECT(p == prog);
+}
+
+TEST_CASE(nms_dynamic_boxes_test)
+{
+    migraphx::program p;
+    auto* mm = p.get_main_module();
+    migraphx::shape sb{migraphx::shape::float_type, {{1, 1, 0}, {6, 20, 0}, {4, 4, 0}}};
+    auto b = mm->add_parameter("boxes", sb);
+    migraphx::shape ss{migraphx::shape::float_type, {{1, 1, 0}, {1, 1, 0}, {6, 20, 0}}};
+    auto s = mm->add_parameter("scores", ss);
+    migraphx::shape smo{migraphx::shape::int64_type, {1}};
+    auto mo = mm->add_parameter("max_output_boxes_per_class", smo);
+    migraphx::shape siou{migraphx::shape::float_type, {1}};
+    auto iou = mm->add_parameter("iou_threshold", siou);
+    migraphx::shape sst{migraphx::shape::float_type, {1}};
+    auto st  = mm->add_parameter("score_threshold", sst);
+    auto ret = mm->add_instruction(
+        migraphx::make_op("nonmaxsuppression", {{"use_dyn_output", true}}), b, s, mo, iou, st);
+    mm->add_return({ret});
+
+    migraphx::onnx_options options;
+    options.default_dyn_dim_value = {6, 20, 0};
+    options.use_dyn_output        = true;
+
+    auto prog = migraphx::parse_onnx("nms_dynamic_boxes_test.onnx", options);
+    EXPECT(p == prog);
+}
+
+TEST_CASE(nms_dynamic_classes_test)
+{
+    migraphx::program p;
+    auto* mm = p.get_main_module();
+    migraphx::shape sb{migraphx::shape::float_type, {1, 6, 4}};
+    auto b = mm->add_parameter("boxes", sb);
+    migraphx::shape ss{migraphx::shape::float_type, {{1, 1, 0}, {1, 10, 0}, {6, 6, 0}}};
+    auto s = mm->add_parameter("scores", ss);
+    migraphx::shape smo{migraphx::shape::int64_type, {1}};
+    auto mo = mm->add_parameter("max_output_boxes_per_class", smo);
+    migraphx::shape siou{migraphx::shape::float_type, {1}};
+    auto iou = mm->add_parameter("iou_threshold", siou);
+    migraphx::shape sst{migraphx::shape::float_type, {1}};
+    auto st  = mm->add_parameter("score_threshold", sst);
+    auto ret = mm->add_instruction(
+        migraphx::make_op("nonmaxsuppression", {{"use_dyn_output", true}}), b, s, mo, iou, st);
+    mm->add_return({ret});
+
+    migraphx::onnx_options options;
+    options.default_dyn_dim_value = {1, 10, 0};
+    options.use_dyn_output        = true;
+
+    auto prog = migraphx::parse_onnx("nms_dynamic_classes_test.onnx", options);
+    EXPECT(p == prog);
+}
+
+TEST_CASE(nms_overwrite_use_dyn_output_test)
+{
+    migraphx::program p;
+    auto* mm = p.get_main_module();
+    migraphx::shape sb{migraphx::shape::float_type, {1, 6, 4}};
+    auto b = mm->add_parameter("boxes", sb);
+
+    migraphx::shape ss{migraphx::shape::float_type, {1, 1, 6}};
+    auto s = mm->add_parameter("scores", ss);
+
+    migraphx::shape smo{migraphx::shape::int64_type, {1}};
+    auto mo = mm->add_parameter("max_output_boxes_per_class", smo);
+
+    migraphx::shape siou{migraphx::shape::float_type, {1}};
+    auto iou = mm->add_parameter("iou_threshold", siou);
+
+    migraphx::shape sst{migraphx::shape::float_type, {1}};
+    auto st = mm->add_parameter("score_threshold", sst);
+
+    auto ret = mm->add_instruction(
+        migraphx::make_op("nonmaxsuppression", {{"use_dyn_output", true}}), b, s, mo, iou, st);
+    mm->add_return({ret});
+
+    migraphx::onnx_options options;
+    options.use_dyn_output = true;
+
+    auto prog = migraphx::parse_onnx("nms_use_dyn_output_false_test.onnx", options);
+    EXPECT(p == prog);
+}
+
 TEST_CASE(nonzero_dynamic_test)
 {
     migraphx::program p;

test/op_shape_test.cpp

@@ -1135,6 +1135,149 @@ TEST_CASE(multinomial)
     throws_shape(migraphx::make_op("multinomial", {{"dtype", dtype}}), s, s);
 }
 
+TEST_CASE(nms_shape)
+{
+    // use_dyn_output == false
+    migraphx::shape boxes_s{migraphx::shape::float_type, {1, 6, 4}};
+    migraphx::shape scores_s{migraphx::shape::float_type, {1, 1, 6}};
+    migraphx::shape max_out_s{migraphx::shape::int64_type, {1}};
+    migraphx::shape iou_thres_s{migraphx::shape::float_type, {1}};
+    migraphx::shape score_thres_s{migraphx::shape::float_type, {1}};
+    migraphx::shape output_s{migraphx::shape::int64_type, {6, 3}};
+    expect_shape(output_s,
+                 migraphx::make_op("nonmaxsuppression",
+                                   {{"center_point_box", true}, {"use_dyn_output", false}}),
+                 boxes_s,
+                 scores_s,
+                 max_out_s,
+                 iou_thres_s,
+                 score_thres_s);
+
+    // use_dyn_output == true
+    output_s = {migraphx::shape::int64_type, {{0, 6, 0}, {3, 3, 0}}};
+    expect_shape(output_s,
+                 migraphx::make_op("nonmaxsuppression",
+                                   {{"center_point_box", true}, {"use_dyn_output", true}}),
+                 boxes_s,
+                 scores_s,
+                 max_out_s,
+                 iou_thres_s,
+                 score_thres_s);
+
+    // dynamic batches
+    boxes_s  = {migraphx::shape::float_type, {{1, 3, 0}, {6, 6, 0}, {4, 4, 0}}};
+    scores_s = {migraphx::shape::float_type, {{1, 3, 0}, {1, 1, 0}, {6, 6, 0}}};
+    output_s = {migraphx::shape::int64_type, {{0, 18, 0}, {3, 3, 0}}};
+    expect_shape(output_s,
+                 migraphx::make_op("nonmaxsuppression",
+                                   {{"center_point_box", true}, {"use_dyn_output", true}}),
+                 boxes_s,
+                 scores_s,
+                 max_out_s,
+                 iou_thres_s,
+                 score_thres_s);
+
+    // dynamic num boxes
+    boxes_s  = {migraphx::shape::float_type, {{1, 1, 0}, {6, 20, 0}, {4, 4, 0}}};
+    scores_s = {migraphx::shape::float_type, {{1, 1, 0}, {1, 1, 0}, {6, 20, 0}}};
+    output_s = {migraphx::shape::int64_type, {{0, 20, 0}, {3, 3, 0}}};
+    expect_shape(output_s,
+                 migraphx::make_op("nonmaxsuppression",
+                                   {{"center_point_box", true}, {"use_dyn_output", true}}),
+                 boxes_s,
+                 scores_s,
+                 max_out_s,
+                 iou_thres_s,
+                 score_thres_s);
+
+    // use_dyn_output false with dynamic input shape
+    throws_shape(migraphx::make_op("nonmaxsuppression",
+                                   {{"center_point_box", true}, {"use_dyn_output", false}}),
+                 boxes_s,
+                 scores_s,
+                 max_out_s,
+                 iou_thres_s,
+                 score_thres_s);
+
+    // dynamic classes
+    boxes_s  = {migraphx::shape::float_type, {{1, 1, 0}, {6, 6, 0}, {4, 4, 0}}};
+    scores_s = {migraphx::shape::float_type, {{1, 1, 0}, {1, 3, 0}, {6, 6, 0}}};
+    output_s = {migraphx::shape::int64_type, {{0, 6, 0}, {3, 3, 0}}};
+    expect_shape(output_s,
+                 migraphx::make_op("nonmaxsuppression",
+                                   {{"center_point_box", true}, {"use_dyn_output", true}}),
+                 boxes_s,
+                 scores_s,
+                 max_out_s,
+                 iou_thres_s,
+                 score_thres_s);
+
+    // fixed mismatch batches
+    boxes_s  = {migraphx::shape::float_type, {2, 6, 4}};
+    scores_s = {migraphx::shape::float_type, {1, 1, 6}};
+    throws_shape(migraphx::make_op("nonmaxsuppression",
+                                   {{"center_point_box", true}, {"use_dyn_output", true}}),
+                 boxes_s,
+                 scores_s,
+                 max_out_s,
+                 iou_thres_s,
+                 score_thres_s);
+
+    // fixed mismatch num boxes
+    boxes_s  = {migraphx::shape::float_type, {1, 6, 4}};
+    scores_s = {migraphx::shape::float_type, {1, 1, 4}};
+    throws_shape(migraphx::make_op("nonmaxsuppression",
+                                   {{"center_point_box", true}, {"use_dyn_output", true}}),
+                 boxes_s,
+                 scores_s,
+                 max_out_s,
+                 iou_thres_s,
+                 score_thres_s);
+
+    // dynamic mismatch batches
+    boxes_s  = {migraphx::shape::float_type, {{1, 4, 0}, {6, 6, 0}, {4, 4, 0}}};
+    scores_s = {migraphx::shape::float_type, {{2, 8, 0}, {1, 1, 0}, {6, 6, 0}}};
+    throws_shape(migraphx::make_op("nonmaxsuppression",
+                                   {{"center_point_box", true}, {"use_dyn_output", true}}),
+                 boxes_s,
+                 scores_s,
+                 max_out_s,
+                 iou_thres_s,
+                 score_thres_s);
+
+    // dynamic mismatch num boxes
+    boxes_s  = {migraphx::shape::float_type, {{1, 1, 0}, {6, 8, 0}, {4, 4, 0}}};
+    scores_s = {migraphx::shape::float_type, {{1, 1, 0}, {1, 1, 0}, {3, 9, 0}}};
+    throws_shape(migraphx::make_op("nonmaxsuppression",
+                                   {{"center_point_box", true}, {"use_dyn_output", true}}),
+                 boxes_s,
+                 scores_s,
+                 max_out_s,
+                 iou_thres_s,
+                 score_thres_s);
+
+    // dynamic number of classes, fixed boxes_s, mismatch batches
+    boxes_s  = {migraphx::shape::float_type, {1, 6, 4}};
+    scores_s = {migraphx::shape::float_type, {{1, 3, 0}, {1, 3, 0}, {6, 6, 0}}};
+    throws_shape(migraphx::make_op("nonmaxsuppression",
+                                   {{"center_point_box", true}, {"use_dyn_output", true}}),
+                 boxes_s,
+                 scores_s,
+                 max_out_s,
+                 iou_thres_s,
+                 score_thres_s);
+    // dynamic number of classes, fixed boxes_s, mismatch num boxes
+    boxes_s  = {migraphx::shape::float_type, {1, 6, 4}};
+    scores_s = {migraphx::shape::float_type, {{1, 1, 0}, {1, 3, 0}, {4, 8, 0}}};
+    throws_shape(migraphx::make_op("nonmaxsuppression",
+                                   {{"center_point_box", true}, {"use_dyn_output", true}}),
+                 boxes_s,
+                 scores_s,
+                 max_out_s,
+                 iou_thres_s,
+                 score_thres_s);
+}
+
 TEST_CASE(pooling_shape)
 {
     migraphx::shape output{migraphx::shape::float_type, {4, 3, 1, 1}};

test/ref_ops_test.cpp

@@ -3624,6 +3624,174 @@ TEST_CASE(neg_test)
     EXPECT(migraphx::verify_range(result_vector, gold));
 }
 
+TEST_CASE(nms_dynamic_out_test)
+{
+    migraphx::program p;
+    auto* mm = p.get_main_module();
+    migraphx::shape boxes_s{migraphx::shape::float_type, {1, 6, 4}};
+    std::vector<float> boxes_vec = {0.5, 0.5,  1.0, 1.0, 0.5, 0.6,  1.0, 1.0, 0.5, 0.4,   1.0, 1.0,
+                                    0.5, 10.5, 1.0, 1.0, 0.5, 10.6, 1.0, 1.0, 0.5, 100.5, 1.0, 1.0};
+
+    migraphx::shape scores_s{migraphx::shape::float_type, {1, 1, 6}};
+    std::vector<float> scores_vec = {0.9, 0.75, 0.6, 0.95, 0.5, 0.3};
+
+    auto boxes_l         = mm->add_literal(migraphx::literal(boxes_s, boxes_vec));
+    auto scores_l        = mm->add_literal(migraphx::literal(scores_s, scores_vec));
+    auto max_out_l       = mm->add_literal(int64_t{4});
+    auto iou_threshold   = mm->add_literal(0.5f);
+    auto score_threshold = mm->add_literal(0.0f);
+
+    auto r = mm->add_instruction(
+        migraphx::make_op("nonmaxsuppression",
+                          {{"center_point_box", true}, {"use_dyn_output", true}}),
+        boxes_l,
+        scores_l,
+        max_out_l,
+        iou_threshold,
+        score_threshold);
+    mm->add_return({r});
+
+    p.compile(migraphx::ref::target{});
+    auto output = p.eval({}).back();
+    std::vector<int64_t> result;
+    output.visit([&](auto out) { result.assign(out.begin(), out.end()); });
+    std::vector<int64_t> gold = {0, 0, 3, 0, 0, 0, 0, 0, 5};
+    EXPECT(migraphx::verify_range(result, gold));
+}
+
+TEST_CASE(nms_dynamic_batch_test)
+{
+    migraphx::program p;
+    auto* mm = p.get_main_module();
+    migraphx::shape boxes_s{migraphx::shape::float_type, {{1, 3, 0}, {6, 6, 0}, {4, 4, 0}}};
+
+    migraphx::shape scores_s{migraphx::shape::float_type, {{1, 3, 0}, {1, 1, 0}, {6, 6, 0}}};
+
+    auto boxes_p         = mm->add_parameter("boxes", boxes_s);
+    auto scores_p        = mm->add_parameter("scores", scores_s);
+    auto max_out_l       = mm->add_literal(int64_t{4});
+    auto iou_threshold   = mm->add_literal(0.5f);
+    auto score_threshold = mm->add_literal(0.0f);
+
+    auto r = mm->add_instruction(
+        migraphx::make_op("nonmaxsuppression",
+                          {{"center_point_box", true}, {"use_dyn_output", true}}),
+        boxes_p,
+        scores_p,
+        max_out_l,
+        iou_threshold,
+        score_threshold);
+    mm->add_return({r});
+
+    p.compile(migraphx::ref::target{});
+
+    std::vector<float> boxes_vec  = {0.5, 0.5,  1.0, 1.0, 0.5, 0.6,  1.0, 1.0, 0.5, 0.4,   1.0, 1.0,
+                                    0.5, 10.5, 1.0, 1.0, 0.5, 10.6, 1.0, 1.0, 0.5, 100.5, 1.0, 1.0,
+                                    0.5, 0.5,  1.0, 1.0, 0.5, 0.6,  1.0, 1.0, 0.5, 0.4,   1.0, 1.0,
+                                    0.5, 10.5, 1.0, 1.0, 0.5, 10.6, 1.0, 1.0, 0.5, 100.5, 1.0, 1.0};
+    std::vector<float> scores_vec = {
+        0.9, 0.75, 0.6, 0.95, 0.5, 0.3, 0.9, 0.75, 0.6, 0.95, 0.5, 0.3};
+
+    migraphx::shape input_fixed_shape0{migraphx::shape::float_type, {2, 6, 4}};
+    migraphx::shape input_fixed_shape1{migraphx::shape::float_type, {2, 1, 6}};
+    migraphx::parameter_map params0;
+    params0["boxes"]  = migraphx::argument(input_fixed_shape0, boxes_vec.data());
+    params0["scores"] = migraphx::argument(input_fixed_shape1, scores_vec.data());
+    auto output       = p.eval(params0).back();
+
+    std::vector<int64_t> result;
+    output.visit([&](auto out) { result.assign(out.begin(), out.end()); });
+    std::vector<int64_t> gold = {0, 0, 3, 0, 0, 0, 0, 0, 5, 1, 0, 3, 1, 0, 0, 1, 0, 5};
+    EXPECT(migraphx::verify_range(result, gold));
+}
+
+TEST_CASE(nms_dynamic_boxes_test)
+{
+    migraphx::program p;
+    auto* mm = p.get_main_module();
+    migraphx::shape boxes_s{migraphx::shape::float_type, {{1, 1, 0}, {4, 20, 0}, {4, 4, 0}}};
+
+    migraphx::shape scores_s{migraphx::shape::float_type, {{1, 1, 0}, {1, 1, 0}, {4, 20, 0}}};
+
+    auto boxes_p         = mm->add_parameter("boxes", boxes_s);
+    auto scores_p        = mm->add_parameter("scores", scores_s);
+    auto max_out_l       = mm->add_literal(int64_t{4});
+    auto iou_threshold   = mm->add_literal(0.5f);
+    auto score_threshold = mm->add_literal(0.0f);
+
+    auto r = mm->add_instruction(
+        migraphx::make_op("nonmaxsuppression",
+                          {{"center_point_box", true}, {"use_dyn_output", true}}),
+        boxes_p,
+        scores_p,
+        max_out_l,
+        iou_threshold,
+        score_threshold);
+    mm->add_return({r});
+
+    p.compile(migraphx::ref::target{});
+
+    std::vector<float> boxes_vec  = {0.5, 0.5,  1.0, 1.0, 0.5, 0.6,  1.0, 1.0, 0.5, 0.4,   1.0, 1.0,
+                                    0.5, 10.5, 1.0, 1.0, 0.5, 10.6, 1.0, 1.0, 0.5, 100.5, 1.0, 1.0};
+    std::vector<float> scores_vec = {0.9, 0.75, 0.6, 0.95, 0.5, 0.3};
+
+    migraphx::shape input_fixed_shape0{migraphx::shape::float_type, {1, 6, 4}};
+    migraphx::shape input_fixed_shape1{migraphx::shape::float_type, {1, 1, 6}};
+    migraphx::parameter_map params0;
+    params0["boxes"]  = migraphx::argument(input_fixed_shape0, boxes_vec.data());
+    params0["scores"] = migraphx::argument(input_fixed_shape1, scores_vec.data());
+    auto output       = p.eval(params0).back();
+
+    std::vector<int64_t> result;
+    output.visit([&](auto out) { result.assign(out.begin(), out.end()); });
+    std::vector<int64_t> gold = {0, 0, 3, 0, 0, 0, 0, 0, 5};
+    EXPECT(migraphx::verify_range(result, gold));
+}
+
+TEST_CASE(nms_dynamic_classes_test)
+{
+    migraphx::program p;
+    auto* mm = p.get_main_module();
+    migraphx::shape boxes_s{migraphx::shape::float_type, {{1, 1, 0}, {6, 6, 0}, {4, 4, 0}}};
+
+    migraphx::shape scores_s{migraphx::shape::float_type, {{1, 1, 0}, {1, 3, 0}, {6, 6, 0}}};
+
+    auto boxes_p         = mm->add_parameter("boxes", boxes_s);
+    auto scores_p        = mm->add_parameter("scores", scores_s);
+    auto max_out_l       = mm->add_literal(int64_t{2});
+    auto iou_threshold   = mm->add_literal(0.5f);
+    auto score_threshold = mm->add_literal(0.0f);
+
+    auto r = mm->add_instruction(
+        migraphx::make_op("nonmaxsuppression",
+                          {{"center_point_box", true}, {"use_dyn_output", true}}),
+        boxes_p,
+        scores_p,
+        max_out_l,
+        iou_threshold,
+        score_threshold);
+    mm->add_return({r});
+
+    p.compile(migraphx::ref::target{});
+
+    std::vector<float> boxes_vec  = {0.0, 0.0,  1.0, 1.0,  0.0, 0.1,   1.0, 1.1,
+                                    0.0, -0.1, 1.0, 0.9,  0.0, 10.0,  1.0, 11.0,
+                                    0.0, 10.1, 1.0, 11.1, 0.0, 100.0, 1.0, 101.0};
+    std::vector<float> scores_vec = {
+        0.9, 0.75, 0.6, 0.95, 0.5, 0.3, 0.9, 0.75, 0.6, 0.95, 0.5, 0.3};
+    migraphx::shape input_fixed_shape0{migraphx::shape::float_type, {1, 6, 4}};
+    migraphx::shape input_fixed_shape1{migraphx::shape::float_type, {1, 2, 6}};
+    migraphx::parameter_map params0;
+    params0["boxes"]  = migraphx::argument(input_fixed_shape0, boxes_vec.data());
+    params0["scores"] = migraphx::argument(input_fixed_shape1, scores_vec.data());
+    auto output       = p.eval(params0).back();
+
+    std::vector<int64_t> result;
+    output.visit([&](auto out) { result.assign(out.begin(), out.end()); });
+    std::vector<int64_t> gold = {0, 0, 3, 0, 0, 0, 0, 1, 3, 0, 1, 0};
+    EXPECT(migraphx::verify_range(result, gold));
+}
+
 TEST_CASE(nms_not_center_test)
 {
     migraphx::program p;
@@ -3642,12 +3810,14 @@ TEST_CASE(nms_not_center_test)
     auto iou_threshold   = mm->add_literal(0.5f);
     auto score_threshold = mm->add_literal(0.0f);
 
-    auto r = mm->add_instruction(migraphx::make_op("nonmaxsuppression"),
-                                 boxes_l,
-                                 scores_l,
-                                 max_out_l,
-                                 iou_threshold,
-                                 score_threshold);
+    // set use_dyn_output back to false in operator map
+    auto r =
+        mm->add_instruction(migraphx::make_op("nonmaxsuppression", {{"use_dyn_output", false}}),
+                            boxes_l,
+                            scores_l,
+                            max_out_l,
+                            iou_threshold,
+                            score_threshold);
     mm->add_return({r});
 
     p.compile(migraphx::ref::target{});
@@ -3675,12 +3845,13 @@ TEST_CASE(nms_test)
     auto iou_threshold   = mm->add_literal(0.5f);
     auto score_threshold = mm->add_literal(0.0f);
 
-    auto r = mm->add_instruction(migraphx::make_op("nonmaxsuppression", {{"center_point_box", 1}}),
-                                 boxes_l,
-                                 scores_l,
-                                 max_out_l,
-                                 iou_threshold,
-                                 score_threshold);
+    auto r =
+        mm->add_instruction(migraphx::make_op("nonmaxsuppression", {{"center_point_box", true}}),
+                            boxes_l,
+                            scores_l,
+                            max_out_l,
+                            iou_threshold,
+                            score_threshold);
     mm->add_return({r});
 
     p.compile(migraphx::ref::target{});
@@ -3712,12 +3883,13 @@ TEST_CASE(nms_transpose1_test)
 
     auto transpose_boxes = mm->add_instruction(
         migraphx::make_op("transpose", {{"permutation", {0, 2, 1}}}), t_boxes_l);
-    auto r = mm->add_instruction(migraphx::make_op("nonmaxsuppression", {{"center_point_box", 1}}),
-                                 transpose_boxes,
-                                 scores_l,
-                                 max_out_l,
-                                 iou_threshold,
-                                 score_threshold);
+    auto r =
+        mm->add_instruction(migraphx::make_op("nonmaxsuppression", {{"center_point_box", true}}),
+                            transpose_boxes,
+                            scores_l,
+                            max_out_l,
+                            iou_threshold,
+                            score_threshold);
     mm->add_return({r});
 
     p.compile(migraphx::ref::target{});
@@ -3749,12 +3921,13 @@ TEST_CASE(nms_transpose2_test)
 
     auto transpose_boxes = mm->add_instruction(
         migraphx::make_op("transpose", {{"permutation", {1, 2, 0}}}), t_boxes_l);
-    auto r = mm->add_instruction(migraphx::make_op("nonmaxsuppression", {{"center_point_box", 1}}),
-                                 transpose_boxes,
-                                 scores_l,
-                                 max_out_l,
-                                 iou_threshold,
-                                 score_threshold);
+    auto r =
+        mm->add_instruction(migraphx::make_op("nonmaxsuppression", {{"center_point_box", true}}),
+                            transpose_boxes,
+                            scores_l,
+                            max_out_l,
+                            iou_threshold,
+                            score_threshold);
     mm->add_return({r});
 
     p.compile(migraphx::ref::target{});

test/simplify_reshapes_test.cpp

@@ -39,6 +39,15 @@ void run_pass(migraphx::module& m)
     migraphx::run_passes(m, {migraphx::simplify_reshapes{}, migraphx::dead_code_elimination{}});
 }
 
+inline std::vector<std::vector<std::size_t>> to_lens(const std::vector<migraphx::shape>& shapes)
+{
+    std::vector<std::vector<std::size_t>> result;
+    std::transform(shapes.begin(), shapes.end(), std::back_inserter(result), [&](const auto& s) {
+        return s.lens();
+    });
+    return result;
+}
+
 TEST_CASE(double_contig)
 {
     migraphx::program p;
@@ -1275,4 +1284,82 @@ TEST_CASE(transpose_slice_single_transpose)
     EXPECT(m1 == m2);
 }
 
+TEST_CASE(transpose_slice_non_packed_axis)
+{
+    migraphx::module m1;
+    {
+        auto x = m1.add_parameter("x", {migraphx::shape::float_type, {2, 384, 36, 64}});
+        auto transpose =
+            m1.add_instruction(migraphx::make_op("transpose", {{"permutation", {0, 2, 1, 3}}}), x);
+        auto slice = m1.add_instruction(
+            migraphx::make_op("slice", {{"axes", {1}}, {"starts", {0}}, {"ends", {12}}}),
+            transpose);
+        auto sqrt = m1.add_instruction(migraphx::make_op("sqrt"), slice);
+        m1.add_return({sqrt});
+    }
+    auto output_shapes = m1.get_output_shapes();
+    run_pass(m1);
+    EXPECT(m1.get_output_shapes() == output_shapes);
+    migraphx::module m2;
+    {
+        auto x = m2.add_parameter("x", {migraphx::shape::float_type, {2, 384, 36, 64}});
+        auto unsqueeze =
+            m2.add_instruction(migraphx::make_op("unsqueeze", {{"axes", {2}}, {"steps", {12}}}), x);
+        auto transpose = m2.add_instruction(
+            migraphx::make_op("transpose", {{"permutation", {3, 0, 2, 1, 4}}}), unsqueeze);
+        auto slice = m2.add_instruction(
+            migraphx::make_op("slice", {{"axes", {0}}, {"starts", {0}}, {"ends", {1}}}), transpose);
+        auto squeeze = m2.add_instruction(migraphx::make_op("squeeze", {{"axes", {0}}}), slice);
+        auto sqrt    = m2.add_instruction(migraphx::make_op("sqrt"), squeeze);
+        m2.add_return({sqrt});
+    }
+    EXPECT(m1 == m2);
+}
+
+TEST_CASE(transpose_slice_non_packed_multi_axis)
+{
+    migraphx::module m1;
+    {
+        auto x = m1.add_parameter("x", {migraphx::shape::float_type, {2, 384, 36, 64}});
+        auto transpose =
+            m1.add_instruction(migraphx::make_op("transpose", {{"permutation", {0, 2, 1, 3}}}), x);
+        auto slice1 = m1.add_instruction(
+            migraphx::make_op("slice", {{"axes", {1}}, {"starts", {0}}, {"ends", {12}}}),
+            transpose);
+        auto slice2 = m1.add_instruction(
+            migraphx::make_op("slice", {{"axes", {1}}, {"starts", {12}}, {"ends", {24}}}),
+            transpose);
+        auto transpose2 = m1.add_instruction(
+            migraphx::make_op("transpose", {{"permutation", {0, 1, 3, 2}}}), slice2);
+        auto slice3 = m1.add_instruction(
+            migraphx::make_op("slice", {{"axes", {1}}, {"starts", {24}}, {"ends", {36}}}),
+            transpose);
+        m1.add_return({slice1, transpose2, slice3});
+    }
+    auto output_shapes = m1.get_output_shapes();
+    run_pass(m1);
+    EXPECT(to_lens(m1.get_output_shapes()) == to_lens(output_shapes));
+    migraphx::module m2;
+    {
+        auto x = m2.add_parameter("x", {migraphx::shape::float_type, {2, 384, 36, 64}});
+        auto unsqueeze =
+            m2.add_instruction(migraphx::make_op("unsqueeze", {{"axes", {2}}, {"steps", {12}}}), x);
+        auto transpose = m2.add_instruction(
+            migraphx::make_op("transpose", {{"permutation", {3, 0, 2, 1, 4}}}), unsqueeze);
+        auto slice1 = m2.add_instruction(
+            migraphx::make_op("slice", {{"axes", {0}}, {"starts", {0}}, {"ends", {1}}}), transpose);
+        auto squeeze1 = m2.add_instruction(migraphx::make_op("squeeze", {{"axes", {0}}}), slice1);
+        auto slice2   = m2.add_instruction(
+            migraphx::make_op("slice", {{"axes", {0}}, {"starts", {1}}, {"ends", {2}}}), transpose);
+        auto squeeze2   = m2.add_instruction(migraphx::make_op("squeeze", {{"axes", {0}}}), slice2);
+        auto transpose2 = m2.add_instruction(
+            migraphx::make_op("transpose", {{"permutation", {0, 1, 3, 2}}}), squeeze2);
+        auto slice3 = m2.add_instruction(
+            migraphx::make_op("slice", {{"axes", {0}}, {"starts", {2}}, {"ends", {3}}}), transpose);
+        auto squeeze3 = m2.add_instruction(migraphx::make_op("squeeze", {{"axes", {0}}}), slice3);
+        m2.add_return({squeeze1, transpose2, squeeze3});
+    }
+    EXPECT(m1.sort() == m2.sort());
+}
+
 int main(int argc, const char* argv[]) { test::run(argc, argv); }

test/verify/test_layernorm.cpp

@@ -68,7 +68,7 @@ struct test_layernorm : verify_program<test_layernorm>
     {
         migraphx::program p;
         auto* mm                 = p.get_main_module();
-        std::vector<size_t> dims = {1, 1, 5};
+        std::vector<size_t> dims = {1, 2, 5};
         auto x = mm->add_parameter("x", migraphx::shape{migraphx::shape::float_type, dims});
         add_layernorm(*mm, x, dims);
         return p;