New `DetectMultiBackend()` class (#5549)

* New `DetectMultiBackend()` class

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* pb to pt fix

* Cleanup

* explicit apply_classifier path

* Cleanup2

* Cleanup3

* Cleanup4

* Cleanup5

* Cleanup6

* val.py MultiBackend inference

* warmup fix

* to device fix

* pt fix

* device fix

* Val cleanup

* COCO128 URL to assets

* half fix

* detect fix

* detect fix 2

* remove half from DetectMultiBackend

* training half handling

* training half handling 2

* training half handling 3

* Cleanup

* Fix CI error

* Add torchscript _extra_files

* Add TorchScript

* Add CoreML

* CoreML cleanup

* New `DetectMultiBackend()` class

* pb to pt fix

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Cleanup

* explicit apply_classifier path

* Cleanup2

* Cleanup3

* Cleanup4

* Cleanup5

* Cleanup6

* val.py MultiBackend inference

* warmup fix

* to device fix

* pt fix

* device fix

* Val cleanup

* COCO128 URL to assets

* half fix

* detect fix

* detect fix 2

* remove half from DetectMultiBackend

* training half handling

* training half handling 2

* training half handling 3

* Cleanup

* Fix CI error

* Add torchscript _extra_files

* Add TorchScript

* Add CoreML

* CoreML cleanup

* revert default to pt

* Add Usage examples

* Cleanup val

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

data/coco128.yaml

@@ -27,4 +27,4 @@ names: ['person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 't
 
 
 # Download script/URL (optional)
-download: https://github.com/ultralytics/yolov5/releases/download/v1.0/coco128.zip
+download: https://ultralytics.com/assets/coco128.zip

detect.py

@@ -14,12 +14,10 @@ Usage:
 
 import argparse
 import os
-import platform
 import sys
 from pathlib import Path
 
 import cv2
-import numpy as np
 import torch
 import torch.backends.cudnn as cudnn
 
@@ -29,13 +27,12 @@ if str(ROOT) not in sys.path:
     sys.path.append(str(ROOT))  # add ROOT to PATH
 ROOT = Path(os.path.relpath(ROOT, Path.cwd()))  # relative
 
-from models.experimental import attempt_load
+from models.common import DetectMultiBackend
 from utils.datasets import IMG_FORMATS, VID_FORMATS, LoadImages, LoadStreams
-from utils.general import (LOGGER, apply_classifier, check_file, check_img_size, check_imshow, check_requirements,
-                           check_suffix, colorstr, increment_path, non_max_suppression, print_args, scale_coords,
-                           strip_optimizer, xyxy2xywh)
+from utils.general import (LOGGER, check_file, check_img_size, check_imshow, check_requirements, colorstr,
+                           increment_path, non_max_suppression, print_args, scale_coords, strip_optimizer, xyxy2xywh)
 from utils.plots import Annotator, colors, save_one_box
-from utils.torch_utils import load_classifier, select_device, time_sync
+from utils.torch_utils import select_device, time_sync
 
 
 @torch.no_grad()
@@ -77,120 +74,45 @@ def run(weights=ROOT / 'yolov5s.pt',  # model.pt path(s)
     save_dir = increment_path(Path(project) / name, exist_ok=exist_ok)  # increment run
     (save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True)  # make dir
 
-    # Initialize
+    # Load model
     device = select_device(device)
-    half &= device.type != 'cpu'  # half precision only supported on CUDA
+    model = DetectMultiBackend(weights, device=device, dnn=dnn)
+    stride, names, pt, jit, onnx = model.stride, model.names, model.pt, model.jit, model.onnx
+    imgsz = check_img_size(imgsz, s=stride)  # check image size
 
-    # Load model
-    w = str(weights[0] if isinstance(weights, list) else weights)
-    classify, suffix, suffixes = False, Path(w).suffix.lower(), ['.pt', '.onnx', '.tflite', '.pb', '']
-    check_suffix(w, suffixes)  # check weights have acceptable suffix
-    pt, onnx, tflite, pb, saved_model = (suffix == x for x in suffixes)  # backend booleans
-    stride, names = 64, [f'class{i}' for i in range(1000)]  # assign defaults
+    # Half
+    half &= pt and device.type != 'cpu'  # half precision only supported by PyTorch on CUDA
     if pt:
-        model = torch.jit.load(w) if 'torchscript' in w else attempt_load(weights, map_location=device)
-        stride = int(model.stride.max())  # model stride
-        names = model.module.names if hasattr(model, 'module') else model.names  # get class names
-        if half:
-            model.half()  # to FP16
-        if classify:  # second-stage classifier
-            modelc = load_classifier(name='resnet50', n=2)  # initialize
-            modelc.load_state_dict(torch.load('resnet50.pt', map_location=device)['model']).to(device).eval()
-    elif onnx:
-        if dnn:
-            check_requirements(('opencv-python>=4.5.4',))
-            net = cv2.dnn.readNetFromONNX(w)
-        else:
-            check_requirements(('onnx', 'onnxruntime-gpu' if torch.has_cuda else 'onnxruntime'))
-            import onnxruntime
-            session = onnxruntime.InferenceSession(w, None)
-    else:  # TensorFlow models
-        import tensorflow as tf
-        if pb:  # https://www.tensorflow.org/guide/migrate#a_graphpb_or_graphpbtxt
-            def wrap_frozen_graph(gd, inputs, outputs):
-                x = tf.compat.v1.wrap_function(lambda: tf.compat.v1.import_graph_def(gd, name=""), [])  # wrapped import
-                return x.prune(tf.nest.map_structure(x.graph.as_graph_element, inputs),
-                               tf.nest.map_structure(x.graph.as_graph_element, outputs))
-
-            graph_def = tf.Graph().as_graph_def()
-            graph_def.ParseFromString(open(w, 'rb').read())
-            frozen_func = wrap_frozen_graph(gd=graph_def, inputs="x:0", outputs="Identity:0")
-        elif saved_model:
-            model = tf.keras.models.load_model(w)
-        elif tflite:
-            if "edgetpu" in w:  # https://www.tensorflow.org/lite/guide/python#install_tensorflow_lite_for_python
-                import tflite_runtime.interpreter as tflri
-                delegate = {'Linux': 'libedgetpu.so.1',  # install libedgetpu https://coral.ai/software/#edgetpu-runtime
-                            'Darwin': 'libedgetpu.1.dylib',
-                            'Windows': 'edgetpu.dll'}[platform.system()]
-                interpreter = tflri.Interpreter(model_path=w, experimental_delegates=[tflri.load_delegate(delegate)])
-            else:
-                interpreter = tf.lite.Interpreter(model_path=w)  # load TFLite model
-            interpreter.allocate_tensors()  # allocate
-            input_details = interpreter.get_input_details()  # inputs
-            output_details = interpreter.get_output_details()  # outputs
-            int8 = input_details[0]['dtype'] == np.uint8  # is TFLite quantized uint8 model
-    imgsz = check_img_size(imgsz, s=stride)  # check image size
+        model.model.half() if half else model.model.float()
 
     # Dataloader
     if webcam:
         view_img = check_imshow()
         cudnn.benchmark = True  # set True to speed up constant image size inference
-        dataset = LoadStreams(source, img_size=imgsz, stride=stride, auto=pt)
+        dataset = LoadStreams(source, img_size=imgsz, stride=stride, auto=pt and not jit)
         bs = len(dataset)  # batch_size
     else:
-        dataset = LoadImages(source, img_size=imgsz, stride=stride, auto=pt)
+        dataset = LoadImages(source, img_size=imgsz, stride=stride, auto=pt and not jit)
         bs = 1  # batch_size
     vid_path, vid_writer = [None] * bs, [None] * bs
 
     # Run inference
     if pt and device.type != 'cpu':
-        model(torch.zeros(1, 3, *imgsz).to(device).type_as(next(model.parameters())))  # run once
+        model(torch.zeros(1, 3, *imgsz).to(device).type_as(next(model.model.parameters())))  # warmup
     dt, seen = [0.0, 0.0, 0.0], 0
-    for path, img, im0s, vid_cap, s in dataset:
+    for path, im, im0s, vid_cap, s in dataset:
         t1 = time_sync()
-        if onnx:
-            img = img.astype('float32')
-        else:
-            img = torch.from_numpy(img).to(device)
-            img = img.half() if half else img.float()  # uint8 to fp16/32
-        img /= 255  # 0 - 255 to 0.0 - 1.0
-        if len(img.shape) == 3:
-            img = img[None]  # expand for batch dim
+        im = torch.from_numpy(im).to(device)
+        im = im.half() if half else im.float()  # uint8 to fp16/32
+        im /= 255  # 0 - 255 to 0.0 - 1.0
+        if len(im.shape) == 3:
+            im = im[None]  # expand for batch dim
         t2 = time_sync()
         dt[0] += t2 - t1
 
         # Inference
-        if pt:
-            visualize = increment_path(save_dir / Path(path).stem, mkdir=True) if visualize else False
-            pred = model(img, augment=augment, visualize=visualize)[0]
-        elif onnx:
-            if dnn:
-                net.setInput(img)
-                pred = torch.tensor(net.forward())
-            else:
-                pred = torch.tensor(session.run([session.get_outputs()[0].name], {session.get_inputs()[0].name: img}))
-        else:  # tensorflow model (tflite, pb, saved_model)
-            imn = img.permute(0, 2, 3, 1).cpu().numpy()  # image in numpy
-            if pb:
-                pred = frozen_func(x=tf.constant(imn)).numpy()
-            elif saved_model:
-                pred = model(imn, training=False).numpy()
-            elif tflite:
-                if int8:
-                    scale, zero_point = input_details[0]['quantization']
-                    imn = (imn / scale + zero_point).astype(np.uint8)  # de-scale
-                interpreter.set_tensor(input_details[0]['index'], imn)
-                interpreter.invoke()
-                pred = interpreter.get_tensor(output_details[0]['index'])
-                if int8:
-                    scale, zero_point = output_details[0]['quantization']
-                    pred = (pred.astype(np.float32) - zero_point) * scale  # re-scale
-            pred[..., 0] *= imgsz[1]  # x
-            pred[..., 1] *= imgsz[0]  # y
-            pred[..., 2] *= imgsz[1]  # w
-            pred[..., 3] *= imgsz[0]  # h
-            pred = torch.tensor(pred)
+        visualize = increment_path(save_dir / Path(path).stem, mkdir=True) if visualize else False
+        pred = model(im, augment=augment, visualize=visualize)
         t3 = time_sync()
         dt[1] += t3 - t2
 
@@ -199,8 +121,7 @@ def run(weights=ROOT / 'yolov5s.pt',  # model.pt path(s)
         dt[2] += time_sync() - t3
 
         # Second-stage classifier (optional)
-        if classify:
-            pred = apply_classifier(pred, modelc, img, im0s)
+        # pred = utils.general.apply_classifier(pred, classifier_model, im, im0s)
 
         # Process predictions
         for i, det in enumerate(pred):  # per image
@@ -212,15 +133,15 @@ def run(weights=ROOT / 'yolov5s.pt',  # model.pt path(s)
                 p, im0, frame = path, im0s.copy(), getattr(dataset, 'frame', 0)
 
             p = Path(p)  # to Path
-            save_path = str(save_dir / p.name)  # img.jpg
-            txt_path = str(save_dir / 'labels' / p.stem) + ('' if dataset.mode == 'image' else f'_{frame}')  # img.txt
-            s += '%gx%g ' % img.shape[2:]  # print string
+            save_path = str(save_dir / p.name)  # im.jpg
+            txt_path = str(save_dir / 'labels' / p.stem) + ('' if dataset.mode == 'image' else f'_{frame}')  # im.txt
+            s += '%gx%g ' % im.shape[2:]  # print string
             gn = torch.tensor(im0.shape)[[1, 0, 1, 0]]  # normalization gain whwh
             imc = im0.copy() if save_crop else im0  # for save_crop
             annotator = Annotator(im0, line_width=line_thickness, example=str(names))
             if len(det):
                 # Rescale boxes from img_size to im0 size
-                det[:, :4] = scale_coords(img.shape[2:], det[:, :4], im0.shape).round()
+                det[:, :4] = scale_coords(im.shape[2:], det[:, :4], im0.shape).round()
 
                 # Print results
                 for c in det[:, -1].unique():

export.py

@@ -21,6 +21,7 @@ TensorFlow.js:
 """
 
 import argparse
+import json
 import os
 import subprocess
 import sys
@@ -54,7 +55,9 @@ def export_torchscript(model, im, file, optimize, prefix=colorstr('TorchScript:'
         f = file.with_suffix('.torchscript.pt')
 
         ts = torch.jit.trace(model, im, strict=False)
-        (optimize_for_mobile(ts) if optimize else ts).save(f)
+        d = {"shape": im.shape, "stride": int(max(model.stride)), "names": model.names}
+        extra_files = {'config.txt': json.dumps(d)}  # torch._C.ExtraFilesMap()
+        (optimize_for_mobile(ts) if optimize else ts).save(f, _extra_files=extra_files)
 
         LOGGER.info(f'{prefix} export success, saved as {f} ({file_size(f):.1f} MB)')
     except Exception as e:

models/common.py

@@ -3,11 +3,14 @@
 Common modules
 """
 
+import json
 import math
+import platform
 import warnings
 from copy import copy
 from pathlib import Path
 
+import cv2
 import numpy as np
 import pandas as pd
 import requests
@@ -17,7 +20,8 @@ from PIL import Image
 from torch.cuda import amp
 
 from utils.datasets import exif_transpose, letterbox
-from utils.general import LOGGER, colorstr, increment_path, make_divisible, non_max_suppression, scale_coords, xyxy2xywh
+from utils.general import (LOGGER, check_requirements, check_suffix, colorstr, increment_path, make_divisible,
+                           non_max_suppression, scale_coords, xywh2xyxy, xyxy2xywh)
 from utils.plots import Annotator, colors, save_one_box
 from utils.torch_utils import time_sync
 
@@ -269,6 +273,128 @@ class Concat(nn.Module):
         return torch.cat(x, self.d)
 
 
+class DetectMultiBackend(nn.Module):
+    # YOLOv5 MultiBackend class for python inference on various backends
+    def __init__(self, weights='yolov5s.pt', device=None, dnn=True):
+        # Usage:
+        #   PyTorch:      weights = *.pt
+        #   TorchScript:            *.torchscript.pt
+        #   CoreML:                 *.mlmodel
+        #   TensorFlow:             *_saved_model
+        #   TensorFlow:             *.pb
+        #   TensorFlow Lite:        *.tflite
+        #   ONNX Runtime:           *.onnx
+        #   OpenCV DNN:             *.onnx with dnn=True
+        super().__init__()
+        w = str(weights[0] if isinstance(weights, list) else weights)
+        suffix, suffixes = Path(w).suffix.lower(), ['.pt', '.onnx', '.tflite', '.pb', '', '.mlmodel']
+        check_suffix(w, suffixes)  # check weights have acceptable suffix
+        pt, onnx, tflite, pb, saved_model, coreml = (suffix == x for x in suffixes)  # backend booleans
+        jit = pt and 'torchscript' in w.lower()
+        stride, names = 64, [f'class{i}' for i in range(1000)]  # assign defaults
+
+        if jit:  # TorchScript
+            LOGGER.info(f'Loading {w} for TorchScript inference...')
+            extra_files = {'config.txt': ''}  # model metadata
+            model = torch.jit.load(w, _extra_files=extra_files)
+            if extra_files['config.txt']:
+                d = json.loads(extra_files['config.txt'])  # extra_files dict
+                stride, names = int(d['stride']), d['names']
+        elif pt:  # PyTorch
+            from models.experimental import attempt_load  # scoped to avoid circular import
+            model = torch.jit.load(w) if 'torchscript' in w else attempt_load(weights, map_location=device)
+            stride = int(model.stride.max())  # model stride
+            names = model.module.names if hasattr(model, 'module') else model.names  # get class names
+        elif coreml:  # CoreML *.mlmodel
+            import coremltools as ct
+            model = ct.models.MLModel(w)
+        elif dnn:  # ONNX OpenCV DNN
+            LOGGER.info(f'Loading {w} for ONNX OpenCV DNN inference...')
+            check_requirements(('opencv-python>=4.5.4',))
+            net = cv2.dnn.readNetFromONNX(w)
+        elif onnx:  # ONNX Runtime
+            LOGGER.info(f'Loading {w} for ONNX Runtime inference...')
+            check_requirements(('onnx', 'onnxruntime-gpu' if torch.has_cuda else 'onnxruntime'))
+            import onnxruntime
+            session = onnxruntime.InferenceSession(w, None)
+        else:  # TensorFlow model (TFLite, pb, saved_model)
+            import tensorflow as tf
+            if pb:  # https://www.tensorflow.org/guide/migrate#a_graphpb_or_graphpbtxt
+                def wrap_frozen_graph(gd, inputs, outputs):
+                    x = tf.compat.v1.wrap_function(lambda: tf.compat.v1.import_graph_def(gd, name=""), [])  # wrapped
+                    return x.prune(tf.nest.map_structure(x.graph.as_graph_element, inputs),
+                                   tf.nest.map_structure(x.graph.as_graph_element, outputs))
+
+                LOGGER.info(f'Loading {w} for TensorFlow *.pb inference...')
+                graph_def = tf.Graph().as_graph_def()
+                graph_def.ParseFromString(open(w, 'rb').read())
+                frozen_func = wrap_frozen_graph(gd=graph_def, inputs="x:0", outputs="Identity:0")
+            elif saved_model:
+                LOGGER.info(f'Loading {w} for TensorFlow saved_model inference...')
+                model = tf.keras.models.load_model(w)
+            elif tflite:  # https://www.tensorflow.org/lite/guide/python#install_tensorflow_lite_for_python
+                if 'edgetpu' in w.lower():
+                    LOGGER.info(f'Loading {w} for TensorFlow Edge TPU inference...')
+                    import tflite_runtime.interpreter as tfli
+                    delegate = {'Linux': 'libedgetpu.so.1',  # install https://coral.ai/software/#edgetpu-runtime
+                                'Darwin': 'libedgetpu.1.dylib',
+                                'Windows': 'edgetpu.dll'}[platform.system()]
+                    interpreter = tfli.Interpreter(model_path=w, experimental_delegates=[tfli.load_delegate(delegate)])
+                else:
+                    LOGGER.info(f'Loading {w} for TensorFlow Lite inference...')
+                    interpreter = tf.lite.Interpreter(model_path=w)  # load TFLite model
+                interpreter.allocate_tensors()  # allocate
+                input_details = interpreter.get_input_details()  # inputs
+                output_details = interpreter.get_output_details()  # outputs
+        self.__dict__.update(locals())  # assign all variables to self
+
+    def forward(self, im, augment=False, visualize=False, val=False):
+        # YOLOv5 MultiBackend inference
+        b, ch, h, w = im.shape  # batch, channel, height, width
+        if self.pt:  # PyTorch
+            y = self.model(im) if self.jit else self.model(im, augment=augment, visualize=visualize)
+            return y if val else y[0]
+        elif self.coreml:  # CoreML *.mlmodel
+            im = im.permute(0, 2, 3, 1).cpu().numpy()  # torch BCHW to numpy BHWC shape(1,320,192,3)
+            im = Image.fromarray((im[0] * 255).astype('uint8'))
+            # im = im.resize((192, 320), Image.ANTIALIAS)
+            y = self.model.predict({'image': im})  # coordinates are xywh normalized
+            box = xywh2xyxy(y['coordinates'] * [[w, h, w, h]])  # xyxy pixels
+            conf, cls = y['confidence'].max(1), y['confidence'].argmax(1).astype(np.float)
+            y = np.concatenate((box, conf.reshape(-1, 1), cls.reshape(-1, 1)), 1)
+        elif self.onnx:  # ONNX
+            im = im.cpu().numpy()  # torch to numpy
+            if self.dnn:  # ONNX OpenCV DNN
+                self.net.setInput(im)
+                y = self.net.forward()
+            else:  # ONNX Runtime
+                y = self.session.run([self.session.get_outputs()[0].name], {self.session.get_inputs()[0].name: im})[0]
+        else:  # TensorFlow model (TFLite, pb, saved_model)
+            im = im.permute(0, 2, 3, 1).cpu().numpy()  # torch BCHW to numpy BHWC shape(1,320,192,3)
+            if self.pb:
+                y = self.frozen_func(x=self.tf.constant(im)).numpy()
+            elif self.saved_model:
+                y = self.model(im, training=False).numpy()
+            elif self.tflite:
+                input, output = self.input_details[0], self.output_details[0]
+                int8 = input['dtype'] == np.uint8  # is TFLite quantized uint8 model
+                if int8:
+                    scale, zero_point = input['quantization']
+                    im = (im / scale + zero_point).astype(np.uint8)  # de-scale
+                self.interpreter.set_tensor(input['index'], im)
+                self.interpreter.invoke()
+                y = self.interpreter.get_tensor(output['index'])
+                if int8:
+                    scale, zero_point = output['quantization']
+                    y = (y.astype(np.float32) - zero_point) * scale  # re-scale
+            y[..., 0] *= w  # x
+            y[..., 1] *= h  # y
+            y[..., 2] *= w  # w
+            y[..., 3] *= h  # h
+        y = torch.tensor(y)
+        return (y, []) if val else y
+
+
 class AutoShape(nn.Module):
     # YOLOv5 input-robust model wrapper for passing cv2/np/PIL/torch inputs. Includes preprocessing, inference and NMS
     conf = 0.25  # NMS confidence threshold

utils/general.py

@@ -785,7 +785,8 @@ def print_mutation(results, hyp, save_dir, bucket):
 
 
 def apply_classifier(x, model, img, im0):
-    # Apply a second stage classifier to yolo outputs
+    # Apply a second stage classifier to YOLO outputs
+    # Example model = torchvision.models.__dict__['efficientnet_b0'](pretrained=True).to(device).eval()
     im0 = [im0] if isinstance(im0, np.ndarray) else im0
     for i, d in enumerate(x):  # per image
         if d is not None and len(d):

utils/torch_utils.py

@@ -17,7 +17,6 @@ import torch
 import torch.distributed as dist
 import torch.nn as nn
 import torch.nn.functional as F
-import torchvision
 
 from utils.general import LOGGER
 
@@ -235,25 +234,6 @@ def model_info(model, verbose=False, img_size=640):
     LOGGER.info(f"Model Summary: {len(list(model.modules()))} layers, {n_p} parameters, {n_g} gradients{fs}")
 
 
-def load_classifier(name='resnet101', n=2):
-    # Loads a pretrained model reshaped to n-class output
-    model = torchvision.models.__dict__[name](pretrained=True)
-
-    # ResNet model properties
-    # input_size = [3, 224, 224]
-    # input_space = 'RGB'
-    # input_range = [0, 1]
-    # mean = [0.485, 0.456, 0.406]
-    # std = [0.229, 0.224, 0.225]
-
-    # Reshape output to n classes
-    filters = model.fc.weight.shape[1]
-    model.fc.bias = nn.Parameter(torch.zeros(n), requires_grad=True)
-    model.fc.weight = nn.Parameter(torch.zeros(n, filters), requires_grad=True)
-    model.fc.out_features = n
-    return model
-
-
 def scale_img(img, ratio=1.0, same_shape=False, gs=32):  # img(16,3,256,416)
     # scales img(bs,3,y,x) by ratio constrained to gs-multiple
     if ratio == 1.0:

val.py

@@ -23,10 +23,10 @@ if str(ROOT) not in sys.path:
     sys.path.append(str(ROOT))  # add ROOT to PATH
 ROOT = Path(os.path.relpath(ROOT, Path.cwd()))  # relative
 
-from models.experimental import attempt_load
+from models.common import DetectMultiBackend
 from utils.callbacks import Callbacks
 from utils.datasets import create_dataloader
-from utils.general import (LOGGER, box_iou, check_dataset, check_img_size, check_requirements, check_suffix, check_yaml,
+from utils.general import (LOGGER, box_iou, check_dataset, check_img_size, check_requirements, check_yaml,
                            coco80_to_coco91_class, colorstr, increment_path, non_max_suppression, print_args,
                            scale_coords, xywh2xyxy, xyxy2xywh)
 from utils.metrics import ConfusionMatrix, ap_per_class
@@ -100,6 +100,7 @@ def run(data,
         name='exp',  # save to project/name
         exist_ok=False,  # existing project/name ok, do not increment
         half=True,  # use FP16 half-precision inference
+        dnn=False,  # use OpenCV DNN for ONNX inference
         model=None,
         dataloader=None,
         save_dir=Path(''),
@@ -110,8 +111,10 @@ def run(data,
     # Initialize/load model and set device
     training = model is not None
     if training:  # called by train.py
-        device = next(model.parameters()).device  # get model device
+        device, pt = next(model.parameters()).device, True  # get model device, PyTorch model
 
+        half &= device.type != 'cpu'  # half precision only supported on CUDA
+        model.half() if half else model.float()
     else:  # called directly
         device = select_device(device, batch_size=batch_size)
 
@@ -120,22 +123,21 @@ def run(data,
         (save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True)  # make dir
 
         # Load model
-        check_suffix(weights, '.pt')
-        model = attempt_load(weights, map_location=device)  # load FP32 model
-        gs = max(int(model.stride.max()), 32)  # grid size (max stride)
-        imgsz = check_img_size(imgsz, s=gs)  # check image size
-
-        # Multi-GPU disabled, incompatible with .half() https://github.com/ultralytics/yolov5/issues/99
-        # if device.type != 'cpu' and torch.cuda.device_count() > 1:
-        #     model = nn.DataParallel(model)
+        model = DetectMultiBackend(weights, device=device, dnn=dnn)
+        stride, pt = model.stride, model.pt
+        imgsz = check_img_size(imgsz, s=stride)  # check image size
+        half &= pt and device.type != 'cpu'  # half precision only supported by PyTorch on CUDA
+        if pt:
+            model.model.half() if half else model.model.float()
+        else:
+            half = False
+            batch_size = 1  # export.py models default to batch-size 1
+            device = torch.device('cpu')
+            LOGGER.info(f'Forcing --batch-size 1 square inference shape(1,3,{imgsz},{imgsz}) for non-PyTorch backends')
 
         # Data
         data = check_dataset(data)  # check
 
-    # Half
-    half &= device.type != 'cpu'  # half precision only supported on CUDA
-    model.half() if half else model.float()
-
     # Configure
     model.eval()
     is_coco = isinstance(data.get('val'), str) and data['val'].endswith('coco/val2017.txt')  # COCO dataset
@@ -145,11 +147,11 @@ def run(data,
 
     # Dataloader
     if not training:
-        if device.type != 'cpu':
-            model(torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(next(model.parameters())))  # run once
+        if pt and device.type != 'cpu':
+            model(torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(next(model.model.parameters())))  # warmup
         pad = 0.0 if task == 'speed' else 0.5
         task = task if task in ('train', 'val', 'test') else 'val'  # path to train/val/test images
-        dataloader = create_dataloader(data[task], imgsz, batch_size, gs, single_cls, pad=pad, rect=True,
+        dataloader = create_dataloader(data[task], imgsz, batch_size, stride, single_cls, pad=pad, rect=pt,
                                        prefix=colorstr(f'{task}: '))[0]
 
     seen = 0
@@ -160,32 +162,33 @@ def run(data,
     dt, p, r, f1, mp, mr, map50, map = [0.0, 0.0, 0.0], 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0
     loss = torch.zeros(3, device=device)
     jdict, stats, ap, ap_class = [], [], [], []
-    for batch_i, (img, targets, paths, shapes) in enumerate(tqdm(dataloader, desc=s)):
+    for batch_i, (im, targets, paths, shapes) in enumerate(tqdm(dataloader, desc=s)):
         t1 = time_sync()
-        img = img.to(device, non_blocking=True)
-        img = img.half() if half else img.float()  # uint8 to fp16/32
-        img /= 255  # 0 - 255 to 0.0 - 1.0
-        targets = targets.to(device)
-        nb, _, height, width = img.shape  # batch size, channels, height, width
+        if pt:
+            im = im.to(device, non_blocking=True)
+            targets = targets.to(device)
+        im = im.half() if half else im.float()  # uint8 to fp16/32
+        im /= 255  # 0 - 255 to 0.0 - 1.0
+        nb, _, height, width = im.shape  # batch size, channels, height, width
         t2 = time_sync()
         dt[0] += t2 - t1
 
-        # Run model
-        out, train_out = model(img, augment=augment)  # inference and training outputs
+        # Inference
+        out, train_out = model(im) if training else model(im, augment=augment, val=True)  # inference, loss outputs
         dt[1] += time_sync() - t2
 
-        # Compute loss
+        # Loss
         if compute_loss:
             loss += compute_loss([x.float() for x in train_out], targets)[1]  # box, obj, cls
 
-        # Run NMS
+        # NMS
         targets[:, 2:] *= torch.Tensor([width, height, width, height]).to(device)  # to pixels
         lb = [targets[targets[:, 0] == i, 1:] for i in range(nb)] if save_hybrid else []  # for autolabelling
         t3 = time_sync()
         out = non_max_suppression(out, conf_thres, iou_thres, labels=lb, multi_label=True, agnostic=single_cls)
         dt[2] += time_sync() - t3
 
-        # Statistics per image
+        # Metrics
         for si, pred in enumerate(out):
             labels = targets[targets[:, 0] == si, 1:]
             nl = len(labels)
@@ -202,12 +205,12 @@ def run(data,
             if single_cls:
                 pred[:, 5] = 0
             predn = pred.clone()
-            scale_coords(img[si].shape[1:], predn[:, :4], shape, shapes[si][1])  # native-space pred
+            scale_coords(im[si].shape[1:], predn[:, :4], shape, shapes[si][1])  # native-space pred
 
             # Evaluate
             if nl:
                 tbox = xywh2xyxy(labels[:, 1:5])  # target boxes
-                scale_coords(img[si].shape[1:], tbox, shape, shapes[si][1])  # native-space labels
+                scale_coords(im[si].shape[1:], tbox, shape, shapes[si][1])  # native-space labels
                 labelsn = torch.cat((labels[:, 0:1], tbox), 1)  # native-space labels
                 correct = process_batch(predn, labelsn, iouv)
                 if plots:
@@ -221,16 +224,16 @@ def run(data,
                 save_one_txt(predn, save_conf, shape, file=save_dir / 'labels' / (path.stem + '.txt'))
             if save_json:
                 save_one_json(predn, jdict, path, class_map)  # append to COCO-JSON dictionary
-            callbacks.run('on_val_image_end', pred, predn, path, names, img[si])
+            callbacks.run('on_val_image_end', pred, predn, path, names, im[si])
 
         # Plot images
         if plots and batch_i < 3:
             f = save_dir / f'val_batch{batch_i}_labels.jpg'  # labels
-            Thread(target=plot_images, args=(img, targets, paths, f, names), daemon=True).start()
+            Thread(target=plot_images, args=(im, targets, paths, f, names), daemon=True).start()
             f = save_dir / f'val_batch{batch_i}_pred.jpg'  # predictions
-            Thread(target=plot_images, args=(img, output_to_target(out), paths, f, names), daemon=True).start()
+            Thread(target=plot_images, args=(im, output_to_target(out), paths, f, names), daemon=True).start()
 
-    # Compute statistics
+    # Compute metrics
     stats = [np.concatenate(x, 0) for x in zip(*stats)]  # to numpy
     if len(stats) and stats[0].any():
         p, r, ap, f1, ap_class = ap_per_class(*stats, plot=plots, save_dir=save_dir, names=names)
@@ -318,6 +321,7 @@ def parse_opt():
     parser.add_argument('--name', default='exp', help='save to project/name')
     parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
     parser.add_argument('--half', action='store_true', help='use FP16 half-precision inference')
+    parser.add_argument('--dnn', action='store_true', help='use OpenCV DNN for ONNX inference')
     opt = parser.parse_args()
     opt.data = check_yaml(opt.data)  # check YAML
     opt.save_json |= opt.data.endswith('coco.yaml')