Simplified usage log API

torchvision/csrc/ops/deform_conv2d.cpp

@@ -20,6 +20,7 @@ at::Tensor deform_conv2d(
     int64_t groups,
     int64_t offset_groups,
     bool use_mask) {
+  C10_LOG_API_USAGE_ONCE("torchvision.csrc.ops.deform_conv2d.deform_conv2d");
   static auto op = c10::Dispatcher::singleton()
                        .findSchemaOrThrow("torchvision::deform_conv2d", "")
                        .typed<decltype(deform_conv2d)>();

torchvision/csrc/ops/nms.cpp

@@ -9,6 +9,7 @@ at::Tensor nms(
     const at::Tensor& dets,
     const at::Tensor& scores,
     double iou_threshold) {
+  C10_LOG_API_USAGE_ONCE("torchvision.csrc.ops.nms.nms");
   static auto op = c10::Dispatcher::singleton()
                        .findSchemaOrThrow("torchvision::nms", "")
                        .typed<decltype(nms)>();

torchvision/csrc/ops/ps_roi_align.cpp

@@ -12,6 +12,7 @@ std::tuple<at::Tensor, at::Tensor> ps_roi_align(
     int64_t pooled_height,
     int64_t pooled_width,
     int64_t sampling_ratio) {
+  C10_LOG_API_USAGE_ONCE("torchvision.csrc.ops.ps_roi_align.ps_roi_align");
   static auto op = c10::Dispatcher::singleton()
                        .findSchemaOrThrow("torchvision::ps_roi_align", "")
                        .typed<decltype(ps_roi_align)>();

torchvision/csrc/ops/ps_roi_pool.cpp

@@ -11,6 +11,7 @@ std::tuple<at::Tensor, at::Tensor> ps_roi_pool(
     double spatial_scale,
     int64_t pooled_height,
     int64_t pooled_width) {
+  C10_LOG_API_USAGE_ONCE("torchvision.csrc.ops.ps_roi_pool.ps_roi_pool");
   static auto op = c10::Dispatcher::singleton()
                        .findSchemaOrThrow("torchvision::ps_roi_pool", "")
                        .typed<decltype(ps_roi_pool)>();

torchvision/csrc/ops/roi_align.cpp

@@ -16,6 +16,7 @@ at::Tensor roi_align(
     bool aligned) // The flag for pixel shift
 // along each axis.
 {
+  C10_LOG_API_USAGE_ONCE("torchvision.csrc.ops.roi_align.roi_align");
   static auto op = c10::Dispatcher::singleton()
                        .findSchemaOrThrow("torchvision::roi_align", "")
                        .typed<decltype(roi_align)>();

torchvision/csrc/ops/roi_pool.cpp

@@ -11,6 +11,7 @@ std::tuple<at::Tensor, at::Tensor> roi_pool(
     double spatial_scale,
     int64_t pooled_height,
     int64_t pooled_width) {
+  C10_LOG_API_USAGE_ONCE("torchvision.csrc.ops.roi_pool.roi_pool");
   static auto op = c10::Dispatcher::singleton()
                        .findSchemaOrThrow("torchvision::roi_pool", "")
                        .typed<decltype(roi_pool)>();

torchvision/datasets/vision.py

@@ -35,7 +35,7 @@ def __init__(
         transform: Optional[Callable] = None,
         target_transform: Optional[Callable] = None,
     ) -> None:
-        _log_api_usage_once("datasets", self.__class__.__name__)
+        _log_api_usage_once(self)
         if isinstance(root, torch._six.string_classes):
             root = os.path.expanduser(root)
         self.root = root

torchvision/models/alexnet.py

@@ -18,7 +18,7 @@
 class AlexNet(nn.Module):
     def __init__(self, num_classes: int = 1000, dropout: float = 0.5) -> None:
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
         self.features = nn.Sequential(
             nn.Conv2d(3, 64, kernel_size=11, stride=4, padding=2),
             nn.ReLU(inplace=True),

torchvision/models/densenet.py

@@ -163,7 +163,7 @@ def __init__(
     ) -> None:
 
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
 
         # First convolution
         self.features = nn.Sequential(

torchvision/models/detection/generalized_rcnn.py

@@ -27,7 +27,7 @@ class GeneralizedRCNN(nn.Module):
 
     def __init__(self, backbone: nn.Module, rpn: nn.Module, roi_heads: nn.Module, transform: nn.Module) -> None:
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
         self.transform = transform
         self.backbone = backbone
         self.rpn = rpn

torchvision/models/detection/retinanet.py

@@ -337,7 +337,7 @@ def __init__(
         topk_candidates=1000,
     ):
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
 
         if not hasattr(backbone, "out_channels"):
             raise ValueError(

torchvision/models/detection/ssd.py

@@ -182,7 +182,7 @@ def __init__(
         positive_fraction: float = 0.25,
     ):
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
 
         self.backbone = backbone
 

torchvision/models/detection/ssdlite.py

@@ -120,7 +120,7 @@ def __init__(
         min_depth: int = 16,
     ):
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
 
         assert not backbone[c4_pos].use_res_connect
         self.features = nn.Sequential(

torchvision/models/efficientnet.py

@@ -170,7 +170,7 @@ def __init__(
             norm_layer (Optional[Callable[..., nn.Module]]): Module specifying the normalization layer to use
         """
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
 
         if not inverted_residual_setting:
             raise ValueError("The inverted_residual_setting should not be empty")

torchvision/models/googlenet.py

@@ -39,7 +39,7 @@ def __init__(
         dropout_aux: float = 0.7,
     ) -> None:
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
         if blocks is None:
             blocks = [BasicConv2d, Inception, InceptionAux]
         if init_weights is None:

torchvision/models/inception.py

@@ -37,7 +37,7 @@ def __init__(
         dropout: float = 0.5,
     ) -> None:
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
         if inception_blocks is None:
             inception_blocks = [BasicConv2d, InceptionA, InceptionB, InceptionC, InceptionD, InceptionE, InceptionAux]
         if init_weights is None:

torchvision/models/mnasnet.py

@@ -98,7 +98,7 @@ class MNASNet(torch.nn.Module):
 
     def __init__(self, alpha: float, num_classes: int = 1000, dropout: float = 0.2) -> None:
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
         assert alpha > 0.0
         self.alpha = alpha
         self.num_classes = num_classes

torchvision/models/mobilenetv2.py

@@ -111,7 +111,7 @@ def __init__(
 
         """
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
 
         if block is None:
             block = InvertedResidual

torchvision/models/mobilenetv3.py

@@ -151,7 +151,7 @@ def __init__(
             dropout (float): The droupout probability
         """
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
 
         if not inverted_residual_setting:
             raise ValueError("The inverted_residual_setting should not be empty")

torchvision/models/optical_flow/raft.py

@@ -440,7 +440,7 @@ def __init__(self, *, feature_encoder, context_encoder, corr_block, update_block
                 If ``None`` (default), the flow is upsampled using interpolation.
         """
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
 
         self.feature_encoder = feature_encoder
         self.context_encoder = context_encoder

torchvision/models/regnet.py

@@ -310,7 +310,7 @@ def __init__(
         activation: Optional[Callable[..., nn.Module]] = None,
     ) -> None:
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
 
         if stem_type is None:
             stem_type = SimpleStemIN

torchvision/models/resnet.py

@@ -174,7 +174,7 @@ def __init__(
         norm_layer: Optional[Callable[..., nn.Module]] = None,
     ) -> None:
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
         if norm_layer is None:
             norm_layer = nn.BatchNorm2d
         self._norm_layer = norm_layer

torchvision/models/segmentation/_utils.py

@@ -13,7 +13,7 @@ class _SimpleSegmentationModel(nn.Module):
 
     def __init__(self, backbone: nn.Module, classifier: nn.Module, aux_classifier: Optional[nn.Module] = None) -> None:
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
         self.backbone = backbone
         self.classifier = classifier
         self.aux_classifier = aux_classifier

torchvision/models/segmentation/lraspp.py

@@ -38,7 +38,7 @@ def __init__(
         self, backbone: nn.Module, low_channels: int, high_channels: int, num_classes: int, inter_channels: int = 128
     ) -> None:
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
         self.backbone = backbone
         self.classifier = LRASPPHead(low_channels, high_channels, num_classes, inter_channels)
 

torchvision/models/shufflenetv2.py

@@ -100,7 +100,7 @@ def __init__(
         inverted_residual: Callable[..., nn.Module] = InvertedResidual,
     ) -> None:
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
 
         if len(stages_repeats) != 3:
             raise ValueError("expected stages_repeats as list of 3 positive ints")

torchvision/models/squeezenet.py

@@ -36,7 +36,7 @@ def forward(self, x: torch.Tensor) -> torch.Tensor:
 class SqueezeNet(nn.Module):
     def __init__(self, version: str = "1_0", num_classes: int = 1000, dropout: float = 0.5) -> None:
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
         self.num_classes = num_classes
         if version == "1_0":
             self.features = nn.Sequential(

torchvision/models/vgg.py

@@ -37,7 +37,7 @@ def __init__(
         self, features: nn.Module, num_classes: int = 1000, init_weights: bool = True, dropout: float = 0.5
     ) -> None:
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
         self.features = features
         self.avgpool = nn.AdaptiveAvgPool2d((7, 7))
         self.classifier = nn.Sequential(

torchvision/models/video/resnet.py

@@ -209,7 +209,7 @@ def __init__(
             zero_init_residual (bool, optional): Zero init bottleneck residual BN. Defaults to False.
         """
         super().__init__()
-        _log_api_usage_once("models", self.__class__.__name__)
+        _log_api_usage_once(self)
         self.inplanes = 64
 
         self.stem = stem()

torchvision/ops/boxes.py

@@ -34,7 +34,8 @@ def nms(boxes: Tensor, scores: Tensor, iou_threshold: float) -> Tensor:
         Tensor: int64 tensor with the indices of the elements that have been kept
         by NMS, sorted in decreasing order of scores
     """
-    _log_api_usage_once("ops", "nms")
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(nms)
     _assert_has_ops()
     return torch.ops.torchvision.nms(boxes, scores, iou_threshold)
 
@@ -63,7 +64,8 @@ def batched_nms(
         Tensor: int64 tensor with the indices of the elements that have been kept by NMS, sorted
         in decreasing order of scores
     """
-    _log_api_usage_once("ops", "batched_nms")
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(batched_nms)
     # Benchmarks that drove the following thresholds are at
     # https://github.com/pytorch/vision/issues/1311#issuecomment-781329339
     if boxes.numel() > (4000 if boxes.device.type == "cpu" else 20000) and not torchvision._is_tracing():
@@ -122,7 +124,8 @@ def remove_small_boxes(boxes: Tensor, min_size: float) -> Tensor:
         Tensor[K]: indices of the boxes that have both sides
         larger than min_size
     """
-    _log_api_usage_once("ops", "remove_small_boxes")
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(remove_small_boxes)
     ws, hs = boxes[:, 2] - boxes[:, 0], boxes[:, 3] - boxes[:, 1]
     keep = (ws >= min_size) & (hs >= min_size)
     keep = torch.where(keep)[0]
@@ -141,7 +144,8 @@ def clip_boxes_to_image(boxes: Tensor, size: Tuple[int, int]) -> Tensor:
     Returns:
         Tensor[N, 4]: clipped boxes
     """
-    _log_api_usage_once("ops", "clip_boxes_to_image")
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(clip_boxes_to_image)
     dim = boxes.dim()
     boxes_x = boxes[..., 0::2]
     boxes_y = boxes[..., 1::2]
@@ -181,8 +185,8 @@ def box_convert(boxes: Tensor, in_fmt: str, out_fmt: str) -> Tensor:
     Returns:
         Tensor[N, 4]: Boxes into converted format.
     """
-
-    _log_api_usage_once("ops", "box_convert")
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(box_convert)
     allowed_fmts = ("xyxy", "xywh", "cxcywh")
     if in_fmt not in allowed_fmts or out_fmt not in allowed_fmts:
         raise ValueError("Unsupported Bounding Box Conversions for given in_fmt and out_fmt")
@@ -232,7 +236,8 @@ def box_area(boxes: Tensor) -> Tensor:
     Returns:
         Tensor[N]: the area for each box
     """
-    _log_api_usage_once("ops", "box_area")
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(box_area)
     boxes = _upcast(boxes)
     return (boxes[:, 2] - boxes[:, 0]) * (boxes[:, 3] - boxes[:, 1])
 
@@ -268,7 +273,8 @@ def box_iou(boxes1: Tensor, boxes2: Tensor) -> Tensor:
     Returns:
         Tensor[N, M]: the NxM matrix containing the pairwise IoU values for every element in boxes1 and boxes2
     """
-    _log_api_usage_once("ops", "box_iou")
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(box_iou)
     inter, union = _box_inter_union(boxes1, boxes2)
     iou = inter / union
     return iou
@@ -290,8 +296,8 @@ def generalized_box_iou(boxes1: Tensor, boxes2: Tensor) -> Tensor:
         Tensor[N, M]: the NxM matrix containing the pairwise generalized IoU values
         for every element in boxes1 and boxes2
     """
-
-    _log_api_usage_once("ops", "generalized_box_iou")
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(generalized_box_iou)
     # degenerate boxes gives inf / nan results
     # so do an early check
     assert (boxes1[:, 2:] >= boxes1[:, :2]).all()
@@ -323,7 +329,8 @@ def masks_to_boxes(masks: torch.Tensor) -> torch.Tensor:
     Returns:
         Tensor[N, 4]: bounding boxes
     """
-    _log_api_usage_once("ops", "masks_to_boxes")
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(masks_to_boxes)
     if masks.numel() == 0:
         return torch.zeros((0, 4), device=masks.device, dtype=torch.float)
 

torchvision/ops/deform_conv.py

@@ -60,8 +60,8 @@ def deform_conv2d(
         >>> # returns
         >>>  torch.Size([4, 5, 8, 8])
     """
-
-    _log_api_usage_once("ops", "deform_conv2d")
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(deform_conv2d)
     _assert_has_ops()
     out_channels = weight.shape[0]
 
@@ -124,6 +124,7 @@ def __init__(
         bias: bool = True,
     ):
         super().__init__()
+        _log_api_usage_once(self)
 
         if in_channels % groups != 0:
             raise ValueError("in_channels must be divisible by groups")

torchvision/ops/feature_pyramid_network.py

@@ -77,7 +77,7 @@ def __init__(
         extra_blocks: Optional[ExtraFPNBlock] = None,
     ):
         super().__init__()
-        _log_api_usage_once("ops", self.__class__.__name__)
+        _log_api_usage_once(self)
         self.inner_blocks = nn.ModuleList()
         self.layer_blocks = nn.ModuleList()
         for in_channels in in_channels_list:

torchvision/ops/focal_loss.py

@@ -32,7 +32,8 @@ def sigmoid_focal_loss(
     Returns:
         Loss tensor with the reduction option applied.
     """
-    _log_api_usage_once("ops", "sigmoid_focal_loss")
+    if not torch.jit.is_scripting() and not torch.jit.is_tracing():
+        _log_api_usage_once(sigmoid_focal_loss)
     p = torch.sigmoid(inputs)
     ce_loss = F.binary_cross_entropy_with_logits(inputs, targets, reduction="none")
     p_t = p * targets + (1 - p) * (1 - targets)