diff --git a/README.zh-CN.md b/README.zh-CN.md
index c81feb86c202..69ce9b72d332 100644
--- a/README.zh-CN.md
+++ b/README.zh-CN.md
@@ -8,7 +8,7 @@
-->
-[中文](https://docs.ultralytics.com/zh/) | [한국어](https://docs.ultralytics.com/ko/) | [日本語](https://docs.ultralytics.com/ja/) | [Русский](https://docs.ultralytics.com/ru/) | [Deutsch](https://docs.ultralytics.com/de/) | [Français](https://docs.ultralytics.com/fr/) | [Español](https://docs.ultralytics.com/es/) | [Português](https://docs.ultralytics.com/pt/) | [हिन्दी](https://docs.ultralytics.com/hi/) | [العربية](https://docs.ultralytics.com/ar/)
+[中文](https://docs.ultralytics.com/zh/) | [한국어](https://docs.ultralytics.com/ko/) | [日本語](https://docs.ultralytics.com/ja/) | [Русский](https://docs.ultralytics.com/ru/) | [Deutsch](https://docs.ultralytics.com/de/) | [Français](https://docs.ultralytics.com/fr/) | [Español](https://docs.ultralytics.com/es/) | [Português](https://docs.ultralytics.com/pt/) | [Türkçe](https://docs.ultralytics.com/tr/) | [Tiếng Việt](https://docs.ultralytics.com/vi/) | [हिन्दी](https://docs.ultralytics.com/hi/) | [العربية](https://docs.ultralytics.com/ar/)
diff --git a/export.py b/export.py
index 8fe7ce8fdce1..214d903c2998 100644
--- a/export.py
+++ b/export.py
@@ -346,6 +346,7 @@ def export_engine(model, im, file, half, dynamic, simplify, workspace=4, verbose
onnx = file.with_suffix(".onnx")
LOGGER.info(f"\n{prefix} starting export with TensorRT {trt.__version__}...")
+ is_trt10 = int(trt.__version__.split(".")[0]) >= 10 # is TensorRT >= 10
assert onnx.exists(), f"failed to export ONNX file: {onnx}"
f = file.with_suffix(".engine") # TensorRT engine file
logger = trt.Logger(trt.Logger.INFO)
@@ -354,9 +355,10 @@ def export_engine(model, im, file, half, dynamic, simplify, workspace=4, verbose
builder = trt.Builder(logger)
config = builder.create_builder_config()
- config.max_workspace_size = workspace * 1 << 30
- # config.set_memory_pool_limit(trt.MemoryPoolType.WORKSPACE, workspace << 30) # fix TRT 8.4 deprecation notice
-
+ if is_trt10:
+ config.set_memory_pool_limit(trt.MemoryPoolType.WORKSPACE, workspace << 30)
+ else: # TensorRT versions 7, 8
+ config.max_workspace_size = workspace * 1 << 30
flag = 1 << int(trt.NetworkDefinitionCreationFlag.EXPLICIT_BATCH)
network = builder.create_network(flag)
parser = trt.OnnxParser(network, logger)
@@ -381,8 +383,10 @@ def export_engine(model, im, file, half, dynamic, simplify, workspace=4, verbose
LOGGER.info(f"{prefix} building FP{16 if builder.platform_has_fast_fp16 and half else 32} engine as {f}")
if builder.platform_has_fast_fp16 and half:
config.set_flag(trt.BuilderFlag.FP16)
- with builder.build_engine(network, config) as engine, open(f, "wb") as t:
- t.write(engine.serialize())
+
+ build = builder.build_serialized_network if is_trt10 else builder.build_engine
+ with build(network, config) as engine, open(f, "wb") as t:
+ t.write(engine if is_trt10 else engine.serialize())
return f, None
@@ -405,7 +409,8 @@ def export_saved_model(
try:
import tensorflow as tf
except Exception:
- check_requirements(f"tensorflow{'' if torch.cuda.is_available() else '-macos' if MACOS else '-cpu'}")
+ check_requirements(f"tensorflow{'' if torch.cuda.is_available() else '-macos' if MACOS else '-cpu'}<=2.15.1")
+
import tensorflow as tf
from tensorflow.python.framework.convert_to_constants import convert_variables_to_constants_v2
diff --git a/models/common.py b/models/common.py
index fd8c998149f5..12244fd4b3cf 100644
--- a/models/common.py
+++ b/models/common.py
@@ -57,8 +57,12 @@
from utils.torch_utils import copy_attr, smart_inference_mode
-def autopad(k, p=None, d=1): # kernel, padding, dilation
- # Pad to 'same' shape outputs
+def autopad(k, p=None, d=1):
+ """
+ Pads kernel to 'same' output shape, adjusting for optional dilation; returns padding size.
+
+ `k`: kernel, `p`: padding, `d`: dilation.
+ """
if d > 1:
k = d * (k - 1) + 1 if isinstance(k, int) else [d * (x - 1) + 1 for x in k] # actual kernel-size
if p is None:
@@ -88,13 +92,19 @@ def forward_fuse(self, x):
class DWConv(Conv):
# Depth-wise convolution
- def __init__(self, c1, c2, k=1, s=1, d=1, act=True): # ch_in, ch_out, kernel, stride, dilation, activation
+ def __init__(self, c1, c2, k=1, s=1, d=1, act=True):
+ """Initializes a depth-wise convolution layer with optional activation; args: input channels (c1), output
+ channels (c2), kernel size (k), stride (s), dilation (d), and activation flag (act).
+ """
super().__init__(c1, c2, k, s, g=math.gcd(c1, c2), d=d, act=act)
class DWConvTranspose2d(nn.ConvTranspose2d):
# Depth-wise transpose convolution
- def __init__(self, c1, c2, k=1, s=1, p1=0, p2=0): # ch_in, ch_out, kernel, stride, padding, padding_out
+ def __init__(self, c1, c2, k=1, s=1, p1=0, p2=0):
+ """Initializes a depth-wise transpose convolutional layer for YOLOv5; args: input channels (c1), output channels
+ (c2), kernel size (k), stride (s), input padding (p1), output padding (p2).
+ """
super().__init__(c1, c2, k, s, p1, p2, groups=math.gcd(c1, c2))
@@ -148,7 +158,10 @@ def forward(self, x):
class Bottleneck(nn.Module):
# Standard bottleneck
- def __init__(self, c1, c2, shortcut=True, g=1, e=0.5): # ch_in, ch_out, shortcut, groups, expansion
+ def __init__(self, c1, c2, shortcut=True, g=1, e=0.5):
+ """Initializes a standard bottleneck layer with optional shortcut and group convolution, supporting channel
+ expansion.
+ """
super().__init__()
c_ = int(c2 * e) # hidden channels
self.cv1 = Conv(c1, c_, 1, 1)
@@ -164,7 +177,10 @@ def forward(self, x):
class BottleneckCSP(nn.Module):
# CSP Bottleneck https://github.com/WongKinYiu/CrossStagePartialNetworks
- def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5): # ch_in, ch_out, number, shortcut, groups, expansion
+ def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5):
+ """Initializes CSP bottleneck with optional shortcuts; args: ch_in, ch_out, number of repeats, shortcut bool,
+ groups, expansion.
+ """
super().__init__()
c_ = int(c2 * e) # hidden channels
self.cv1 = Conv(c1, c_, 1, 1)
@@ -206,7 +222,10 @@ def forward(self, x):
class C3(nn.Module):
# CSP Bottleneck with 3 convolutions
- def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5): # ch_in, ch_out, number, shortcut, groups, expansion
+ def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5):
+ """Initializes C3 module with options for channel count, bottleneck repetition, shortcut usage, group
+ convolutions, and expansion.
+ """
super().__init__()
c_ = int(c2 * e) # hidden channels
self.cv1 = Conv(c1, c_, 1, 1)
@@ -283,7 +302,13 @@ def forward(self, x):
class SPPF(nn.Module):
# Spatial Pyramid Pooling - Fast (SPPF) layer for YOLOv5 by Glenn Jocher
- def __init__(self, c1, c2, k=5): # equivalent to SPP(k=(5, 9, 13))
+ def __init__(self, c1, c2, k=5):
+ """
+ Initializes YOLOv5 SPPF layer with given channels and kernel size for YOLOv5 model, combining convolution and
+ max pooling.
+
+ Equivalent to SPP(k=(5, 9, 13)).
+ """
super().__init__()
c_ = c1 // 2 # hidden channels
self.cv1 = Conv(c1, c_, 1, 1)
@@ -302,19 +327,26 @@ def forward(self, x):
class Focus(nn.Module):
# Focus wh information into c-space
- def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True): # ch_in, ch_out, kernel, stride, padding, groups
+ def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True):
+ """Initializes Focus module to concentrate width-height info into channel space with configurable convolution
+ parameters.
+ """
super().__init__()
self.conv = Conv(c1 * 4, c2, k, s, p, g, act=act)
# self.contract = Contract(gain=2)
- def forward(self, x): # x(b,c,w,h) -> y(b,4c,w/2,h/2)
+ def forward(self, x):
+ """Processes input through Focus mechanism, reshaping (b,c,w,h) to (b,4c,w/2,h/2) then applies convolution."""
return self.conv(torch.cat((x[..., ::2, ::2], x[..., 1::2, ::2], x[..., ::2, 1::2], x[..., 1::2, 1::2]), 1))
# return self.conv(self.contract(x))
class GhostConv(nn.Module):
# Ghost Convolution https://github.com/huawei-noah/ghostnet
- def __init__(self, c1, c2, k=1, s=1, g=1, act=True): # ch_in, ch_out, kernel, stride, groups
+ def __init__(self, c1, c2, k=1, s=1, g=1, act=True):
+ """Initializes GhostConv with in/out channels, kernel size, stride, groups, and activation; halves out channels
+ for efficiency.
+ """
super().__init__()
c_ = c2 // 2 # hidden channels
self.cv1 = Conv(c1, c_, k, s, None, g, act=act)
@@ -328,7 +360,8 @@ def forward(self, x):
class GhostBottleneck(nn.Module):
# Ghost Bottleneck https://github.com/huawei-noah/ghostnet
- def __init__(self, c1, c2, k=3, s=1): # ch_in, ch_out, kernel, stride
+ def __init__(self, c1, c2, k=3, s=1):
+ """Initializes GhostBottleneck with ch_in `c1`, ch_out `c2`, kernel size `k`, stride `s`; see https://github.com/huawei-noah/ghostnet."""
super().__init__()
c_ = c2 // 2
self.conv = nn.Sequential(
@@ -494,18 +527,34 @@ def __init__(self, weights="yolov5s.pt", device=torch.device("cpu"), dnn=False,
output_names = []
fp16 = False # default updated below
dynamic = False
- for i in range(model.num_bindings):
- name = model.get_binding_name(i)
- dtype = trt.nptype(model.get_binding_dtype(i))
- if model.binding_is_input(i):
- if -1 in tuple(model.get_binding_shape(i)): # dynamic
- dynamic = True
- context.set_binding_shape(i, tuple(model.get_profile_shape(0, i)[2]))
- if dtype == np.float16:
- fp16 = True
- else: # output
- output_names.append(name)
- shape = tuple(context.get_binding_shape(i))
+ is_trt10 = not hasattr(model, "num_bindings")
+ num = range(model.num_io_tensors) if is_trt10 else range(model.num_bindings)
+ for i in num:
+ if is_trt10:
+ name = model.get_tensor_name(i)
+ dtype = trt.nptype(model.get_tensor_dtype(name))
+ is_input = model.get_tensor_mode(name) == trt.TensorIOMode.INPUT
+ if is_input:
+ if -1 in tuple(model.get_tensor_shape(name)): # dynamic
+ dynamic = True
+ context.set_input_shape(name, tuple(model.get_profile_shape(name, 0)[2]))
+ if dtype == np.float16:
+ fp16 = True
+ else: # output
+ output_names.append(name)
+ shape = tuple(context.get_tensor_shape(name))
+ else:
+ name = model.get_binding_name(i)
+ dtype = trt.nptype(model.get_binding_dtype(i))
+ if model.binding_is_input(i):
+ if -1 in tuple(model.get_binding_shape(i)): # dynamic
+ dynamic = True
+ context.set_binding_shape(i, tuple(model.get_profile_shape(0, i)[2]))
+ if dtype == np.float16:
+ fp16 = True
+ else: # output
+ output_names.append(name)
+ shape = tuple(context.get_binding_shape(i))
im = torch.from_numpy(np.empty(shape, dtype=dtype)).to(device)
bindings[name] = Binding(name, dtype, shape, im, int(im.data_ptr()))
binding_addrs = OrderedDict((n, d.ptr) for n, d in bindings.items())
@@ -982,10 +1031,14 @@ def print(self):
"""Logs the string representation of the current object's state via the LOGGER."""
LOGGER.info(self.__str__())
- def __len__(self): # override len(results)
+ def __len__(self):
+ """Returns the number of results stored, overrides the default len(results)."""
return self.n
- def __str__(self): # override print(results)
+ def __str__(self):
+ """Returns a string representation of the model's results, suitable for printing, overrides default
+ print(results).
+ """
return self._run(pprint=True) # print results
def __repr__(self):
@@ -995,7 +1048,8 @@ def __repr__(self):
class Proto(nn.Module):
# YOLOv5 mask Proto module for segmentation models
- def __init__(self, c1, c_=256, c2=32): # ch_in, number of protos, number of masks
+ def __init__(self, c1, c_=256, c2=32):
+ """Initializes YOLOv5 Proto module for segmentation with input, proto, and mask channels configuration."""
super().__init__()
self.cv1 = Conv(c1, c_, k=3)
self.upsample = nn.Upsample(scale_factor=2, mode="nearest")
diff --git a/models/experimental.py b/models/experimental.py
index 62ee802f6248..6152cef1b389 100644
--- a/models/experimental.py
+++ b/models/experimental.py
@@ -11,8 +11,12 @@
class Sum(nn.Module):
- # Weighted sum of 2 or more layers https://arxiv.org/abs/1911.09070
- def __init__(self, n, weight=False): # n: number of inputs
+ """Weighted sum of 2 or more layers https://arxiv.org/abs/1911.09070."""
+
+ def __init__(self, n, weight=False):
+ """Initializes a module to sum outputs of layers with number of inputs `n` and optional weighting, supporting 2+
+ inputs.
+ """
super().__init__()
self.weight = weight # apply weights boolean
self.iter = range(n - 1) # iter object
@@ -33,8 +37,12 @@ def forward(self, x):
class MixConv2d(nn.Module):
- # Mixed Depth-wise Conv https://arxiv.org/abs/1907.09595
- def __init__(self, c1, c2, k=(1, 3), s=1, equal_ch=True): # ch_in, ch_out, kernel, stride, ch_strategy
+ """Mixed Depth-wise Conv https://arxiv.org/abs/1907.09595."""
+
+ def __init__(self, c1, c2, k=(1, 3), s=1, equal_ch=True):
+ """Initializes MixConv2d with mixed depth-wise convolutional layers, taking input and output channels (c1, c2),
+ kernel sizes (k), stride (s), and channel distribution strategy (equal_ch).
+ """
super().__init__()
n = len(k) # number of convolutions
if equal_ch: # equal c_ per group
diff --git a/models/tf.py b/models/tf.py
index 006a66d2b0f6..2a5cd566c406 100644
--- a/models/tf.py
+++ b/models/tf.py
@@ -190,15 +190,25 @@ def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True, w=None):
super().__init__()
self.conv = TFConv(c1 * 4, c2, k, s, p, g, act, w.conv)
- def call(self, inputs): # x(b,w,h,c) -> y(b,w/2,h/2,4c)
- # inputs = inputs / 255 # normalize 0-255 to 0-1
+ def call(self, inputs):
+ """
+ Performs pixel shuffling and convolution on input tensor, downsampling by 2 and expanding channels by 4.
+
+ Example x(b,w,h,c) -> y(b,w/2,h/2,4c).
+ """
inputs = [inputs[:, ::2, ::2, :], inputs[:, 1::2, ::2, :], inputs[:, ::2, 1::2, :], inputs[:, 1::2, 1::2, :]]
return self.conv(tf.concat(inputs, 3))
class TFBottleneck(keras.layers.Layer):
# Standard bottleneck
- def __init__(self, c1, c2, shortcut=True, g=1, e=0.5, w=None): # ch_in, ch_out, shortcut, groups, expansion
+ def __init__(self, c1, c2, shortcut=True, g=1, e=0.5, w=None):
+ """
+ Initializes a standard bottleneck layer for TensorFlow models, expanding and contracting channels with optional
+ shortcut.
+
+ Arguments are ch_in, ch_out, shortcut, groups, expansion.
+ """
super().__init__()
c_ = int(c2 * e) # hidden channels
self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1)
@@ -364,7 +374,10 @@ def call(self, inputs):
class TFDetect(keras.layers.Layer):
# TF YOLOv5 Detect layer
- def __init__(self, nc=80, anchors=(), ch=(), imgsz=(640, 640), w=None): # detection layer
+ def __init__(self, nc=80, anchors=(), ch=(), imgsz=(640, 640), w=None):
+ """Initializes YOLOv5 detection layer for TensorFlow with configurable classes, anchors, channels, and image
+ size.
+ """
super().__init__()
self.stride = tf.convert_to_tensor(w.stride.numpy(), dtype=tf.float32)
self.nc = nc # number of classes
@@ -454,7 +467,13 @@ def call(self, inputs):
class TFUpsample(keras.layers.Layer):
# TF version of torch.nn.Upsample()
- def __init__(self, size, scale_factor, mode, w=None): # warning: all arguments needed including 'w'
+ def __init__(self, size, scale_factor, mode, w=None):
+ """
+ Initializes a TensorFlow upsampling layer with specified size, scale_factor, and mode, ensuring scale_factor is
+ even.
+
+ Warning: all arguments needed including 'w'
+ """
super().__init__()
assert scale_factor % 2 == 0, "scale_factor must be multiple of 2"
self.upsample = lambda x: tf.image.resize(x, (x.shape[1] * scale_factor, x.shape[2] * scale_factor), mode)
@@ -481,7 +500,8 @@ def call(self, inputs):
return tf.concat(inputs, self.d)
-def parse_model(d, ch, model, imgsz): # model_dict, input_channels(3)
+def parse_model(d, ch, model, imgsz):
+ """Parses a model definition dict `d` to create YOLOv5 model layers, including dynamic channel adjustments."""
LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10} {'module':<40}{'arguments':<30}")
anchors, nc, gd, gw, ch_mul = (
d["anchors"],
@@ -562,7 +582,10 @@ def parse_model(d, ch, model, imgsz): # model_dict, input_channels(3)
class TFModel:
# TF YOLOv5 model
- def __init__(self, cfg="yolov5s.yaml", ch=3, nc=None, model=None, imgsz=(640, 640)): # model, channels, classes
+ def __init__(self, cfg="yolov5s.yaml", ch=3, nc=None, model=None, imgsz=(640, 640)):
+ """Initializes TF YOLOv5 model with specified configuration, channels, classes, model instance, and input
+ size.
+ """
super().__init__()
if isinstance(cfg, dict):
self.yaml = cfg # model dict
@@ -640,7 +663,10 @@ def call(self, input, topk_all, iou_thres, conf_thres):
)
@staticmethod
- def _nms(x, topk_all=100, iou_thres=0.45, conf_thres=0.25): # agnostic NMS
+ def _nms(x, topk_all=100, iou_thres=0.45, conf_thres=0.25):
+ """Performs agnostic non-maximum suppression (NMS) on detected objects, filtering based on IoU and confidence
+ thresholds.
+ """
boxes, classes, scores = x
class_inds = tf.cast(tf.argmax(classes, axis=-1), tf.float32)
scores_inp = tf.reduce_max(scores, -1)
diff --git a/models/yolo.py b/models/yolo.py
index ef6c1015f41e..ca62f934fc3f 100644
--- a/models/yolo.py
+++ b/models/yolo.py
@@ -75,7 +75,8 @@ class Detect(nn.Module):
dynamic = False # force grid reconstruction
export = False # export mode
- def __init__(self, nc=80, anchors=(), ch=(), inplace=True): # detection layer
+ def __init__(self, nc=80, anchors=(), ch=(), inplace=True):
+ """Initializes YOLOv5 detection layer with specified classes, anchors, channels, and inplace operations."""
super().__init__()
self.nc = nc # number of classes
self.no = nc + 5 # number of outputs per anchor
@@ -183,7 +184,8 @@ def _profile_one_layer(self, m, x, dt):
if c:
LOGGER.info(f"{sum(dt):10.2f} {'-':>10s} {'-':>10s} Total")
- def fuse(self): # fuse model Conv2d() + BatchNorm2d() layers
+ def fuse(self):
+ """Fuses Conv2d() and BatchNorm2d() layers in the model to improve inference speed."""
LOGGER.info("Fusing layers... ")
for m in self.model.modules():
if isinstance(m, (Conv, DWConv)) and hasattr(m, "bn"):
@@ -193,7 +195,8 @@ def fuse(self): # fuse model Conv2d() + BatchNorm2d() layers
self.info()
return self
- def info(self, verbose=False, img_size=640): # print model information
+ def info(self, verbose=False, img_size=640):
+ """Prints model information given verbosity and image size, e.g., `info(verbose=True, img_size=640)`."""
model_info(self, verbose, img_size)
def _apply(self, fn):
@@ -212,7 +215,8 @@ def _apply(self, fn):
class DetectionModel(BaseModel):
# YOLOv5 detection model
- def __init__(self, cfg="yolov5s.yaml", ch=3, nc=None, anchors=None): # model, input channels, number of classes
+ def __init__(self, cfg="yolov5s.yaml", ch=3, nc=None, anchors=None):
+ """Initializes YOLOv5 model with configuration file, input channels, number of classes, and custom anchors."""
super().__init__()
if isinstance(cfg, dict):
self.yaml = cfg # model dict
@@ -303,8 +307,12 @@ def _clip_augmented(self, y):
y[-1] = y[-1][:, i:] # small
return y
- def _initialize_biases(self, cf=None): # initialize biases into Detect(), cf is class frequency
- # https://arxiv.org/abs/1708.02002 section 3.3
+ def _initialize_biases(self, cf=None):
+ """
+ Initializes biases for YOLOv5's Detect() module, optionally using class frequencies (cf).
+
+ For details see https://arxiv.org/abs/1708.02002 section 3.3.
+ """
# cf = torch.bincount(torch.tensor(np.concatenate(dataset.labels, 0)[:, 0]).long(), minlength=nc) + 1.
m = self.model[-1] # Detect() module
for mi, s in zip(m.m, m.stride): # from
@@ -328,7 +336,10 @@ def __init__(self, cfg="yolov5s-seg.yaml", ch=3, nc=None, anchors=None):
class ClassificationModel(BaseModel):
# YOLOv5 classification model
- def __init__(self, cfg=None, model=None, nc=1000, cutoff=10): # yaml, model, number of classes, cutoff index
+ def __init__(self, cfg=None, model=None, nc=1000, cutoff=10):
+ """Initializes YOLOv5 model with config file `cfg`, input channels `ch`, number of classes `nc`, and `cuttoff`
+ index.
+ """
super().__init__()
self._from_detection_model(model, nc, cutoff) if model is not None else self._from_yaml(cfg)
@@ -354,8 +365,8 @@ def _from_yaml(self, cfg):
self.model = None
-def parse_model(d, ch): # model_dict, input_channels(3)
- # Parse a YOLOv5 model.yaml dictionary
+def parse_model(d, ch):
+ """Parses a YOLOv5 model from a dict `d`, configuring layers based on input channels `ch` and model architecture."""
LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10} {'module':<40}{'arguments':<30}")
anchors, nc, gd, gw, act, ch_mul = (
d["anchors"],
diff --git a/pyproject.toml b/pyproject.toml
index 26765dfadada..5748b907cf30 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -24,6 +24,7 @@ build-backend = "setuptools.build_meta"
# Project settings -----------------------------------------------------------------------------------------------------
[project]
+version = "7.0.0"
name = "YOLOv5"
description = "Ultralytics YOLOv5 for SOTA object detection, instance segmentation and image classification."
readme = "README.md"
@@ -77,7 +78,7 @@ dependencies = [
"thop>=0.1.1", # FLOPs computation
"pandas>=1.1.4",
"seaborn>=0.11.0", # plotting
- "ultralytics>=8.0.232"
+ "ultralytics>=8.1.47"
]
# Optional dependencies ------------------------------------------------------------------------------------------------
diff --git a/requirements.txt b/requirements.txt
index 545d3c1c2921..3892abe07308 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -6,7 +6,7 @@ gitpython>=3.1.30
matplotlib>=3.3
numpy>=1.23.5
opencv-python>=4.1.1
-Pillow>=9.4.0
+pillow>=10.3.0
psutil # system resources
PyYAML>=5.3.1
requests>=2.23.0
@@ -47,3 +47,4 @@ setuptools>=65.5.1 # Snyk vulnerability fix
# mss # screenshots
# albumentations>=1.0.3
# pycocotools>=2.0.6 # COCO mAP
+wheel>=0.38.0 # not directly required, pinned by Snyk to avoid a vulnerability
diff --git a/segment/train.py b/segment/train.py
index ce59df9c635b..5a6e9afb8ec0 100644
--- a/segment/train.py
+++ b/segment/train.py
@@ -95,7 +95,12 @@
GIT_INFO = check_git_info()
-def train(hyp, opt, device, callbacks): # hyp is path/to/hyp.yaml or hyp dictionary
+def train(hyp, opt, device, callbacks):
+ """
+ Trains the YOLOv5 model on a dataset, managing hyperparameters, model optimization, logging, and validation.
+
+ `hyp` is path/to/hyp.yaml or hyp dictionary.
+ """
(
save_dir,
epochs,
diff --git a/train.py b/train.py
index dfc45d78e5ce..493642adbcc7 100644
--- a/train.py
+++ b/train.py
@@ -100,7 +100,13 @@
GIT_INFO = check_git_info()
-def train(hyp, opt, device, callbacks): # hyp is path/to/hyp.yaml or hyp dictionary
+def train(hyp, opt, device, callbacks):
+ """
+ Trains YOLOv5 model with given hyperparameters, options, and device, managing datasets, model architecture, loss
+ computation, and optimizer steps.
+
+ `hyp` argument is path/to/hyp.yaml or hyp dictionary.
+ """
save_dir, epochs, batch_size, weights, single_cls, evolve, data, cfg, resume, noval, nosave, workers, freeze = (
Path(opt.save_dir),
opt.epochs,
diff --git a/utils/activations.py b/utils/activations.py
index 6218eb58440a..928ae55a0b60 100644
--- a/utils/activations.py
+++ b/utils/activations.py
@@ -59,8 +59,10 @@ def forward(self, x):
class FReLU(nn.Module):
- # FReLU activation https://arxiv.org/abs/2007.11824
+ """FReLU activation https://arxiv.org/abs/2007.11824."""
+
def __init__(self, c1, k=3): # ch_in, kernel
+ """Initializes FReLU activation with channel `c1` and kernel size `k`."""
super().__init__()
self.conv = nn.Conv2d(c1, c1, k, 1, 1, groups=c1, bias=False)
self.bn = nn.BatchNorm2d(c1)
@@ -103,7 +105,8 @@ class MetaAconC(nn.Module):
See "Activate or Not: Learning Customized Activation" https://arxiv.org/pdf/2009.04759.pdf.
"""
- def __init__(self, c1, k=1, s=1, r=16): # ch_in, kernel, stride, r
+ def __init__(self, c1, k=1, s=1, r=16):
+ """Initializes MetaAconC with params: channel_in (c1), kernel size (k=1), stride (s=1), reduction (r=16)."""
super().__init__()
c2 = max(r, c1 // r)
self.p1 = nn.Parameter(torch.randn(1, c1, 1, 1))
diff --git a/utils/augmentations.py b/utils/augmentations.py
index dfac11f28776..872e7259560e 100644
--- a/utils/augmentations.py
+++ b/utils/augmentations.py
@@ -412,8 +412,13 @@ def mixup(im, labels, im2, labels2):
return im, labels
-def box_candidates(box1, box2, wh_thr=2, ar_thr=100, area_thr=0.1, eps=1e-16): # box1(4,n), box2(4,n)
- # Compute candidate boxes: box1 before augment, box2 after augment, wh_thr (pixels), aspect_ratio_thr, area_ratio
+def box_candidates(box1, box2, wh_thr=2, ar_thr=100, area_thr=0.1, eps=1e-16):
+ """
+ Filters bounding box candidates by minimum width-height threshold `wh_thr` (pixels), aspect ratio threshold
+ `ar_thr`, and area ratio threshold `area_thr`.
+
+ box1(4,n) is before augmentation, box2(4,n) is after augmentation.
+ """
w1, h1 = box1[2] - box1[0], box1[3] - box1[1]
w2, h2 = box2[2] - box2[0], box2[3] - box2[1]
ar = np.maximum(w2 / (h2 + eps), h2 / (w2 + eps)) # aspect ratio
@@ -482,7 +487,12 @@ def __init__(self, size=(640, 640), auto=False, stride=32):
self.auto = auto # pass max size integer, automatically solve for short side using stride
self.stride = stride # used with auto
- def __call__(self, im): # im = np.array HWC
+ def __call__(self, im):
+ """
+ Resizes and pads input image `im` (HWC format) to specified dimensions, maintaining aspect ratio.
+
+ im = np.array HWC
+ """
imh, imw = im.shape[:2]
r = min(self.h / imh, self.w / imw) # ratio of new/old
h, w = round(imh * r), round(imw * r) # resized image
@@ -500,7 +510,12 @@ def __init__(self, size=640):
super().__init__()
self.h, self.w = (size, size) if isinstance(size, int) else size
- def __call__(self, im): # im = np.array HWC
+ def __call__(self, im):
+ """
+ Applies center crop to the input image and resizes it to a specified size, maintaining aspect ratio.
+
+ im = np.array HWC
+ """
imh, imw = im.shape[:2]
m = min(imh, imw) # min dimension
top, left = (imh - m) // 2, (imw - m) // 2
@@ -514,7 +529,13 @@ def __init__(self, half=False):
super().__init__()
self.half = half
- def __call__(self, im): # im = np.array HWC in BGR order
+ def __call__(self, im):
+ """
+ Converts BGR np.array image from HWC to RGB CHW format, and normalizes to [0, 1], with support for FP16 if
+ `half=True`.
+
+ im = np.array HWC in BGR order
+ """
im = np.ascontiguousarray(im.transpose((2, 0, 1))[::-1]) # HWC to CHW -> BGR to RGB -> contiguous
im = torch.from_numpy(im) # to torch
im = im.half() if self.half else im.float() # uint8 to fp16/32
diff --git a/utils/dataloaders.py b/utils/dataloaders.py
index 3991563e7910..6dba2c8d897a 100644
--- a/utils/dataloaders.py
+++ b/utils/dataloaders.py
@@ -1069,8 +1069,13 @@ def flatten_recursive(path=DATASETS_DIR / "coco128"):
shutil.copyfile(file, new_path / Path(file).name)
-def extract_boxes(path=DATASETS_DIR / "coco128"): # from utils.dataloaders import *; extract_boxes()
- # Convert detection dataset into classification dataset, with one directory per class
+def extract_boxes(path=DATASETS_DIR / "coco128"):
+ """
+ Converts a detection dataset to a classification dataset, creating a directory for each class and extracting
+ bounding boxes.
+
+ Example: from utils.dataloaders import *; extract_boxes()
+ """
path = Path(path) # images dir
shutil.rmtree(path / "classification") if (path / "classification").is_dir() else None # remove existing
files = list(path.rglob("*.*"))
@@ -1089,7 +1094,7 @@ def extract_boxes(path=DATASETS_DIR / "coco128"): # from utils.dataloaders impo
for j, x in enumerate(lb):
c = int(x[0]) # class
- f = (path / "classifier") / f"{c}" / f"{path.stem}_{im_file.stem}_{j}.jpg" # new filename
+ f = (path / "classification") / f"{c}" / f"{path.stem}_{im_file.stem}_{j}.jpg" # new filename
if not f.parent.is_dir():
f.parent.mkdir(parents=True)
@@ -1262,7 +1267,7 @@ def get_json(self, save=False, verbose=False):
"""Generates dataset JSON for Ultralytics HUB, optionally saves or prints it; save=bool, verbose=bool."""
def _round(labels):
- # Update labels to integer class and 6 decimal place floats
+ """Rounds class labels to integers and coordinates to 4 decimal places for improved label accuracy."""
return [[int(c), *(round(x, 4) for x in points)] for c, *points in labels]
for split in "train", "val", "test":
diff --git a/utils/general.py b/utils/general.py
index 661475354adc..5a9325eec757 100644
--- a/utils/general.py
+++ b/utils/general.py
@@ -351,8 +351,12 @@ def run_once():
return run_once() or run_once() # check twice to increase robustness to intermittent connectivity issues
-def git_describe(path=ROOT): # path must be a directory
- # Return human-readable git description, i.e. v5.0-5-g3e25f1e https://git-scm.com/docs/git-describe
+def git_describe(path=ROOT):
+ """
+ Returns a human-readable git description of the repository at `path`, or an empty string on failure.
+
+ Example output is 'fv5.0-5-g3e25f1e'. See https://git-scm.com/docs/git-describe.
+ """
try:
assert (Path(path) / ".git").is_dir()
return check_output(f"git -C {path} describe --tags --long --always", shell=True).decode()[:-1]
@@ -767,8 +771,12 @@ def labels_to_image_weights(labels, nc=80, class_weights=np.ones(80)):
return (class_weights.reshape(1, nc) * class_counts).sum(1)
-def coco80_to_coco91_class(): # converts 80-index (val2014) to 91-index (paper)
- # https://tech.amikelive.com/node-718/what-object-categories-labels-are-in-coco-dataset/
+def coco80_to_coco91_class():
+ """
+ Converts COCO 80-class index to COCO 91-class index used in the paper.
+
+ Reference: https://tech.amikelive.com/node-718/what-object-categories-labels-are-in-coco-dataset/
+ """
# a = np.loadtxt('data/coco.names', dtype='str', delimiter='\n')
# b = np.loadtxt('data/coco_paper.names', dtype='str', delimiter='\n')
# x1 = [list(a[i] == b).index(True) + 1 for i in range(80)] # darknet to coco
@@ -1108,8 +1116,13 @@ def non_max_suppression(
return output
-def strip_optimizer(f="best.pt", s=""): # from utils.general import *; strip_optimizer()
- # Strip optimizer from 'f' to finalize training, optionally save as 's'
+def strip_optimizer(f="best.pt", s=""):
+ """
+ Strips optimizer and optionally saves checkpoint to finalize training; arguments are file path 'f' and save path
+ 's'.
+
+ Example: from utils.general import *; strip_optimizer()
+ """
x = torch.load(f, map_location=torch.device("cpu"))
if x.get("ema"):
x["model"] = x["ema"] # replace model with ema
diff --git a/utils/google_app_engine/additional_requirements.txt b/utils/google_app_engine/additional_requirements.txt
index c1a2af2c1145..821c3caf3cbf 100644
--- a/utils/google_app_engine/additional_requirements.txt
+++ b/utils/google_app_engine/additional_requirements.txt
@@ -1,5 +1,5 @@
# add these requirements in your app on top of the existing ones
pip==23.3
Flask==2.3.2
-gunicorn==19.10.0
+gunicorn==22.0.0
werkzeug>=3.0.1 # not directly required, pinned by Snyk to avoid a vulnerability
diff --git a/utils/loggers/__init__.py b/utils/loggers/__init__.py
index 2a2c5d734c2e..011ec7c8915b 100644
--- a/utils/loggers/__init__.py
+++ b/utils/loggers/__init__.py
@@ -382,7 +382,7 @@ def __init__(self, opt, console_logger, include=("tb", "wandb", "clearml")):
prefix = colorstr("ClearML: ")
LOGGER.warning(
f"{prefix}WARNING ⚠️ ClearML is installed but not configured, skipping ClearML logging."
- f" See https://github.com/ultralytics/yolov5/tree/master/utils/loggers/clearml#readme"
+ f" See https://docs.ultralytics.com/yolov5/tutorials/clearml_logging_integration"
)
else:
self.clearml = None
diff --git a/utils/loss.py b/utils/loss.py
index 8a910e12ad6f..9d09f9df0261 100644
--- a/utils/loss.py
+++ b/utils/loss.py
@@ -8,8 +8,8 @@
from utils.torch_utils import de_parallel
-def smooth_BCE(eps=0.1): # https://github.com/ultralytics/yolov3/issues/238#issuecomment-598028441
- # return positive, negative label smoothing BCE targets
+def smooth_BCE(eps=0.1):
+ """Returns label smoothing BCE targets for reducing overfitting; pos: `1.0 - 0.5*eps`, neg: `0.5*eps`. For details see https://github.com/ultralytics/yolov3/issues/238#issuecomment-598028441"""
return 1.0 - 0.5 * eps, 0.5 * eps
@@ -132,6 +132,7 @@ def __init__(self, model, autobalance=False):
self.device = device
def __call__(self, p, targets): # predictions, targets
+ """Performs forward pass, calculating class, box, and object loss for given predictions and targets."""
lcls = torch.zeros(1, device=self.device) # class loss
lbox = torch.zeros(1, device=self.device) # box loss
lobj = torch.zeros(1, device=self.device) # object loss
diff --git a/utils/plots.py b/utils/plots.py
index e1b073dfb1ad..cb5edabc6c41 100644
--- a/utils/plots.py
+++ b/utils/plots.py
@@ -67,7 +67,8 @@ def __call__(self, i, bgr=False):
return (c[2], c[1], c[0]) if bgr else c
@staticmethod
- def hex2rgb(h): # rgb order (PIL)
+ def hex2rgb(h):
+ """Converts hexadecimal color `h` to an RGB tuple (PIL-compatible) with order (R, G, B)."""
return tuple(int(h[1 + i : 1 + i + 2], 16) for i in (0, 2, 4))
@@ -225,8 +226,13 @@ def plot_lr_scheduler(optimizer, scheduler, epochs=300, save_dir=""):
plt.close()
-def plot_val_txt(): # from utils.plots import *; plot_val()
- # Plot val.txt histograms
+def plot_val_txt():
+ """
+ Plots 2D and 1D histograms of bounding box centers from 'val.txt' using matplotlib, saving as 'hist2d.png' and
+ 'hist1d.png'.
+
+ Example: from utils.plots import *; plot_val()
+ """
x = np.loadtxt("val.txt", dtype=np.float32)
box = xyxy2xywh(x[:, :4])
cx, cy = box[:, 0], box[:, 1]
@@ -242,8 +248,12 @@ def plot_val_txt(): # from utils.plots import *; plot_val()
plt.savefig("hist1d.png", dpi=200)
-def plot_targets_txt(): # from utils.plots import *; plot_targets_txt()
- # Plot targets.txt histograms
+def plot_targets_txt():
+ """
+ Plots histograms of object detection targets from 'targets.txt', saving the figure as 'targets.jpg'.
+
+ Example: from utils.plots import *; plot_targets_txt()
+ """
x = np.loadtxt("targets.txt", dtype=np.float32).T
s = ["x targets", "y targets", "width targets", "height targets"]
fig, ax = plt.subplots(2, 2, figsize=(8, 8), tight_layout=True)
@@ -255,8 +265,13 @@ def plot_targets_txt(): # from utils.plots import *; plot_targets_txt()
plt.savefig("targets.jpg", dpi=200)
-def plot_val_study(file="", dir="", x=None): # from utils.plots import *; plot_val_study()
- # Plot file=study.txt generated by val.py (or plot all study*.txt in dir)
+def plot_val_study(file="", dir="", x=None):
+ """
+ Plots validation study results from 'study*.txt' files in a directory or a specific file, comparing model
+ performance and speed.
+
+ Example: from utils.plots import *; plot_val_study()
+ """
save_dir = Path(file).parent if file else Path(dir)
plot2 = False # plot additional results
if plot2:
@@ -381,8 +396,12 @@ def imshow_cls(im, labels=None, pred=None, names=None, nmax=25, verbose=False, f
return f
-def plot_evolve(evolve_csv="path/to/evolve.csv"): # from utils.plots import *; plot_evolve()
- # Plot evolve.csv hyp evolution results
+def plot_evolve(evolve_csv="path/to/evolve.csv"):
+ """
+ Plots hyperparameter evolution results from a given CSV, saving the plot and displaying best results.
+
+ Example: from utils.plots import *; plot_evolve()
+ """
evolve_csv = Path(evolve_csv)
data = pd.read_csv(evolve_csv)
keys = [x.strip() for x in data.columns]
diff --git a/utils/segment/loss.py b/utils/segment/loss.py
index 29f1bcbb7e77..fa0c10939b70 100644
--- a/utils/segment/loss.py
+++ b/utils/segment/loss.py
@@ -44,6 +44,7 @@ def __init__(self, model, autobalance=False, overlap=False):
self.device = device
def __call__(self, preds, targets, masks): # predictions, targets, model
+ """Evaluates YOLOv5 model's loss for given predictions, targets, and masks; returns total loss components."""
p, proto = preds
bs, nm, mask_h, mask_w = proto.shape # batch size, number of masks, mask height, mask width
lcls = torch.zeros(1, device=self.device)
diff --git a/utils/torch_utils.py b/utils/torch_utils.py
index c2c760efa404..4929d21cdf83 100644
--- a/utils/torch_utils.py
+++ b/utils/torch_utils.py
@@ -325,7 +325,9 @@ def model_info(model, verbose=False, imgsz=640):
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
+ """Scales an image tensor `img` of shape (bs,3,y,x) by `ratio`, optionally maintaining the original shape, padded to
+ multiples of `gs`.
+ """
if ratio == 1.0:
return img
h, w = img.shape[2:]