DetectMultiBackend() --half handling (ultralytics#6945)

* DetectMultiBackend() `--half` handling

* CI fixes

* rename .half to .fp16 to avoid conflict

* warmup fix

* val update

* engine update

* engine update

detect.py

@@ -89,19 +89,10 @@ def run(weights=ROOT / 'yolov5s.pt',  # model.pt path(s)
 
     # Load model
     device = select_device(device)
-    model = DetectMultiBackend(weights, device=device, dnn=dnn, data=data)
-    stride, names, pt, jit, onnx, engine = model.stride, model.names, model.pt, model.jit, model.onnx, model.engine
+    model = DetectMultiBackend(weights, device=device, dnn=dnn, data=data, fp16=half)
+    stride, names, pt = model.stride, model.names, model.pt
     imgsz = check_img_size(imgsz, s=stride)  # check image size
 
-    # Half
-    half &= (pt or jit or onnx or engine) and device.type != 'cpu'  # FP16 supported on limited backends with CUDA
-    if pt or jit:
-        model.model.half() if half else model.model.float()
-    elif engine and model.trt_fp16_input != half:
-        LOGGER.info('model ' + (
-            'requires' if model.trt_fp16_input else 'incompatible with') + ' --half. Adjusting automatically.')
-        half = model.trt_fp16_input
-
     # Dataloader
     if webcam:
         view_img = check_imshow()
@@ -114,12 +105,12 @@ def run(weights=ROOT / 'yolov5s.pt',  # model.pt path(s)
     vid_path, vid_writer = [None] * bs, [None] * bs
 
     # Run inference
-    model.warmup(imgsz=(1 if pt else bs, 3, *imgsz), half=half)  # warmup
+    model.warmup(imgsz=(1 if pt else bs, 3, *imgsz))  # warmup
     dt, seen = [0.0, 0.0, 0.0], 0
     for path, im, im0s, vid_cap, s in dataset:
         t1 = time_sync()
         im = torch.from_numpy(im).to(device)
-        im = im.half() if half else im.float()  # uint8 to fp16/32
+        im = im.half() if model.fp16 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

models/common.py

@@ -277,7 +277,7 @@ def forward(self, x):
 
 class DetectMultiBackend(nn.Module):
     # YOLOv5 MultiBackend class for python inference on various backends
-    def __init__(self, weights='yolov5s.pt', device=None, dnn=False, data=None):
+    def __init__(self, weights='yolov5s.pt', device=torch.device('cpu'), dnn=False, data=None, fp16=False):
         # Usage:
         #   PyTorch:              weights = *.pt
         #   TorchScript:                    *.torchscript
@@ -297,6 +297,7 @@ def __init__(self, weights='yolov5s.pt', device=None, dnn=False, data=None):
         pt, jit, onnx, xml, engine, coreml, saved_model, pb, tflite, edgetpu, tfjs = self.model_type(w)  # get backend
         stride, names = 64, [f'class{i}' for i in range(1000)]  # assign defaults
         w = attempt_download(w)  # download if not local
+        fp16 &= (pt or jit or onnx or engine) and device.type != 'cpu'  # FP16
         if data:  # data.yaml path (optional)
             with open(data, errors='ignore') as f:
                 names = yaml.safe_load(f)['names']  # class names
@@ -305,11 +306,13 @@ def __init__(self, weights='yolov5s.pt', device=None, dnn=False, data=None):
             model = attempt_load(weights if isinstance(weights, list) else w, map_location=device)
             stride = max(int(model.stride.max()), 32)  # model stride
             names = model.module.names if hasattr(model, 'module') else model.names  # get class names
+            model.half() if fp16 else model.float()
             self.model = model  # explicitly assign for to(), cpu(), cuda(), half()
         elif 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)
+            model.half() if fp16 else model.float()
             if extra_files['config.txt']:
                 d = json.loads(extra_files['config.txt'])  # extra_files dict
                 stride, names = int(d['stride']), d['names']
@@ -338,19 +341,19 @@ def __init__(self, weights='yolov5s.pt', device=None, dnn=False, data=None):
             import tensorrt as trt  # https://developer.nvidia.com/nvidia-tensorrt-download
             check_version(trt.__version__, '7.0.0', hard=True)  # require tensorrt>=7.0.0
             Binding = namedtuple('Binding', ('name', 'dtype', 'shape', 'data', 'ptr'))
-            trt_fp16_input = False
             logger = trt.Logger(trt.Logger.INFO)
             with open(w, 'rb') as f, trt.Runtime(logger) as runtime:
                 model = runtime.deserialize_cuda_engine(f.read())
             bindings = OrderedDict()
+            fp16 = False  # default updated below
             for index in range(model.num_bindings):
                 name = model.get_binding_name(index)
                 dtype = trt.nptype(model.get_binding_dtype(index))
                 shape = tuple(model.get_binding_shape(index))
                 data = torch.from_numpy(np.empty(shape, dtype=np.dtype(dtype))).to(device)
                 bindings[name] = Binding(name, dtype, shape, data, int(data.data_ptr()))
                 if model.binding_is_input(index) and dtype == np.float16:
-                    trt_fp16_input = True
+                    fp16 = True
             binding_addrs = OrderedDict((n, d.ptr) for n, d in bindings.items())
             context = model.create_execution_context()
             batch_size = bindings['images'].shape[0]
@@ -458,11 +461,11 @@ def forward(self, im, augment=False, visualize=False, val=False):
         y = torch.tensor(y) if isinstance(y, np.ndarray) else y
         return (y, []) if val else y
 
-    def warmup(self, imgsz=(1, 3, 640, 640), half=False):
+    def warmup(self, imgsz=(1, 3, 640, 640)):
         # Warmup model by running inference once
         if self.pt or self.jit or self.onnx or self.engine:  # warmup types
             if isinstance(self.device, torch.device) and self.device.type != 'cpu':  # only warmup GPU models
-                im = torch.zeros(*imgsz).to(self.device).type(torch.half if half else torch.float)  # input image
+                im = torch.zeros(*imgsz).to(self.device).type(torch.half if self.fp16 else torch.float)  # input image
                 self.forward(im)  # warmup
 
     @staticmethod

val.py

@@ -125,7 +125,6 @@ def run(data,
     training = model is not None
     if training:  # called by train.py
         device, pt, jit, engine = next(model.parameters()).device, True, False, False  # 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
@@ -136,23 +135,17 @@ def run(data,
         (save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True)  # make dir
 
         # Load model
-        model = DetectMultiBackend(weights, device=device, dnn=dnn, data=data)
-        stride, pt, jit, onnx, engine = model.stride, model.pt, model.jit, model.onnx, model.engine
+        model = DetectMultiBackend(weights, device=device, dnn=dnn, data=data, fp16=half)
+        stride, pt, jit, engine = model.stride, model.pt, model.jit, model.engine
         imgsz = check_img_size(imgsz, s=stride)  # check image size
-        half &= (pt or jit or onnx or engine) and device.type != 'cpu'  # FP16 supported on limited backends with CUDA
-        if pt or jit:
-            model.model.half() if half else model.model.float()
-        elif engine:
+        half = model.fp16  # FP16 supported on limited backends with CUDA
+        if engine:
             batch_size = model.batch_size
-            if model.trt_fp16_input != half:
-                LOGGER.info('model ' + (
-                    'requires' if model.trt_fp16_input else 'incompatible with') + ' --half. Adjusting automatically.')
-                half = model.trt_fp16_input
         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')
+            device = model.device
+            if not pt or jit:
+                batch_size = 1  # export.py models default to batch-size 1
+                LOGGER.info(f'Forcing --batch-size 1 square inference (1,3,{imgsz},{imgsz}) for non-PyTorch models')
 
         # Data
         data = check_dataset(data)  # check
@@ -166,7 +159,7 @@ def run(data,
 
     # Dataloader
     if not training:
-        model.warmup(imgsz=(1 if pt else batch_size, 3, imgsz, imgsz), half=half)  # warmup
+        model.warmup(imgsz=(1 if pt else batch_size, 3, imgsz, imgsz))  # warmup
         pad = 0.0 if task in ('speed', 'benchmark') else 0.5
         rect = False if task == 'benchmark' else pt  # square inference for benchmarks
         task = task if task in ('train', 'val', 'test') else 'val'  # path to train/val/test images