[Fix]: Fix mmrotate and regression script (#1917)

* fix pipeline

* disable test for openvino and sdk+torchscript for mmpose

configs/mmrotate/rotated-detection_onnxruntime_static.py

@@ -1,3 +1,3 @@
 _base_ = ['./rotated-detection_static.py', '../_base_/backends/onnxruntime.py']
 
-onnx_config = dict(output_names=['dets', 'labels'], input_shape=None)
+onnx_config = dict(output_names=['dets', 'labels'], input_shape=[1024, 1024])

mmdeploy/codebase/mmrotate/deploy/rotated_detection.py

@@ -85,12 +85,12 @@ def process_model_config(model_cfg: Config,
         cfg.test_pipeline[0].type = 'mmdet.LoadImageFromNDArray'
 
     pipeline = cfg.test_pipeline
-
-    for i, transform in enumerate(pipeline):
-        # for static exporting
-        if input_shape is not None and transform.type == 'Resize':
-            pipeline[i].keep_ratio = False
-            pipeline[i].scale = tuple(input_shape)
+    # for static exporting
+    if input_shape is not None:
+        for i, transform in enumerate(pipeline):
+            if transform.type in ['Resize', 'mmdet.Resize']:
+                pipeline[i].keep_ratio = False
+                pipeline[i].scale = tuple(input_shape)
 
     pipeline = [
         transform for transform in pipeline
@@ -209,15 +209,12 @@ class RotatedDetection(BaseTask):
         cfg = process_model_config(self.model_cfg, imgs, input_shape)
 
         pipeline = cfg.test_pipeline
+        # for static exporting
         if not dynamic_flag:
-            transform = pipeline[1]
-            if 'transforms' in transform:
-                transform_list = transform['transforms']
-                for i, step in enumerate(transform_list):
-                    if step['type'] == 'Pad' and 'pad_to_square' in step \
-                       and step['pad_to_square']:
-                        transform_list.pop(i)
-                        break
+            for i, trans in enumerate(pipeline):
+                if trans['type'] == 'Pad' and 'pad_to_square' in trans \
+                   and trans['pad_to_square']:
+                    trans.pop(i)
         test_pipeline = Compose(pipeline)
 
         data = []
@@ -261,6 +258,7 @@ class RotatedDetection(BaseTask):
         input_shape = get_input_shape(self.deploy_cfg)
         model_cfg = process_model_config(self.model_cfg, [''], input_shape)
         pipeline = model_cfg.test_pipeline
+        pipeline = replace_RResize(pipeline)
         meta_keys = [
             'filename', 'ori_filename', 'ori_shape', 'img_shape', 'pad_shape',
             'scale_factor', 'flip', 'flip_direction', 'img_norm_cfg',

tests/regression/mmpose.yml

@@ -51,7 +51,7 @@ openvino:
     deploy_config: configs/mmpose/pose-detection_openvino_static-256x192.py
   pipeline_openvino_static_fp32_256x256: &pipeline_openvino_static_fp32_256x256
     convert_image: *convert_image
-    backend_test: *default_backend_test
+    backend_test: False
     deploy_config: configs/mmpose/pose-detection_openvino_static-256x256.py
 
 ncnn:
@@ -74,7 +74,6 @@ torchscript:
   pipeline_ts_static_fp32: &pipeline_ts_fp32
     convert_image: *convert_image
     backend_test: *default_backend_test
-    sdk_config: *sdk_static
     deploy_config: configs/mmpose/pose-detection_torchscript.py
 
 models:

tools/regression_test.py

@@ -332,7 +332,7 @@ def get_pytorch_result(model_name: str, meta_info: dict, checkpoint_path: Path,
     }
 
     # get pytorch fps value
-    fps_info = model_info.get('Metadata').get('inference time (ms/im)')
+    fps_info = model_info.get('Metadata', {}).get('inference time (ms/im)')
     if fps_info is None:
         fps = '-'
     elif isinstance(fps_info, list):