delete onnxruntime suport

image_classification/volo/volo.py

@@ -5,7 +5,6 @@
 import numpy as np
 
 import ailia
-import onnxruntime
 import volo_labels
 
 # import original modules
@@ -33,16 +32,11 @@
 # Arguemnt Parser Config
 # ======================
 parser = get_base_parser('Image classification model: VOLO', IMAGE_PATH, None)
-parser.add_argument(
-    '--onnx',
-    action='store_true',
-    help='execute onnxruntime version.'
-)
 parser.add_argument(
     '--arch',
     default="volo_d1_224", 
     type=str, 
-    choices=["volo_d1_224" ,"volo_d1_384","volo_d2_224", "volo_d2_384","volo_d3_224","volo_d3_448", "volo_d4_224", "volo_d4_448" ,"volo_d5_224"]
+    choices=["volo_d1_224" ,"volo_d1_384","volo_d2_224", "volo_d2_384","volo_d3_224","volo_d3_448", "volo_d4_224", "volo_d4_448" ,"volo_d5_224","volo_d5_448","volo_d5_512"]
 
 )
 args = update_parser(parser)
@@ -55,8 +49,8 @@
 MODEL_PATH  = args.arch + '.onnx.prototxt'
 REMOTE_PATH = 'https://storage.googleapis.com/ailia-models/volo/'
 
-IMAGE_HEIGHT = int(args.arch.split("_")[-1])
-IMAGE_WIDTH  = int(args.arch.split("_")[-1])
+IMAGE_HEIGHT = 224
+IMAGE_WIDTH  = 224
 # ======================
 # Main functions
 # ======================
@@ -86,11 +80,7 @@ def resize_and_center_crop(image, input_size, crop_pct):
 
 def recognize_from_image():
     # net initialize
-    if args.onnx:
-        session = onnxruntime.InferenceSession(WEIGHT_PATH)
-        first_input_name = session.get_inputs()[0].name
-    else:
-        net = ailia.Net(MODEL_PATH, WEIGHT_PATH, env_id=args.env_id)
+    net = ailia.Net(MODEL_PATH, WEIGHT_PATH, env_id=args.env_id)
 
     # input image loop
     for image_path in args.input:
@@ -110,19 +100,13 @@ def recognize_from_image():
             logger.info('BENCHMARK mode')
             for i in range(args.benchmark_count):
                 start = int(round(time.time() * 1000))
-                if args.onnx:
-                    results = session.run([], {first_input_name: input_data})[0]
-                    preds_ailia=np.array(results)
-                else:
-                    preds_ailia = net.predict(input_data)
+
+                preds_ailia = net.predict(input_data)
+
                 end = int(round(time.time() * 1000))
                 logger.info(f'\tailia processing time {end - start} ms')
         else:
-            if args.onnx:
-                results = session.run([], {first_input_name: input_data})[0]
-                preds_ailia=np.array(results)
-            else:
-                preds_ailia = net.predict(input_data)
+            preds_ailia = net.predict(input_data)
 
         # postprocessing
         print_results(preds_ailia, volo_labels.imagenet_category)
@@ -131,12 +115,7 @@ def recognize_from_image():
 
 def recognize_from_video():
     # net initialize
-
-    if args.onnx:
-        session = onnxruntime.InferenceSession(WEIGHT_PATH)
-        first_input_name = session.get_inputs()[0].name
-    else:
-        net = ailia.Net(MODEL_PATH, WEIGHT_PATH, env_id=args.env_id)
+    net = ailia.Net(MODEL_PATH, WEIGHT_PATH, env_id=args.env_id)
 
     capture = webcamera_utils.get_capture(args.video)
 
@@ -162,11 +141,7 @@ def recognize_from_video():
         input_data = input_data.astype(np.float32)
 
         # inference
-        if args.onnx:
-            results = session.run([], {first_input_name: input_data})[0]
-            preds_ailia=np.array(results)
-        else:
-            preds_ailia = net.predict(input_data)
+        preds_ailia = net.predict(input_data)
 
         # postprocessing
         plot_results(frame, preds_ailia, volo_labels.imagenet_category)