`ultralytics 8.0.35` TensorRT, ONNX and OpenVINO predict and val (#929)

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Eric Pedley <ericpedley@gmail.com>
single_channel
Glenn Jocher 2 years ago committed by GitHub
parent d32b339373
commit 977fd8f0b8
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23

@ -198,16 +198,16 @@ See [Classification Docs](https://docs.ultralytics.com/tasks/classification/) fo
<div align="center">
<a href="https://roboflow.com/?ref=ultralytics">
<img src="https://github.com/ultralytics/yolov5/releases/download/v1.0/logo-roboflow.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/partners/logo-roboflow.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/social/logo-transparent.png" width="15%" height="0" alt="" />
<a href="https://cutt.ly/yolov5-readme-clearml">
<img src="https://github.com/ultralytics/yolov5/releases/download/v1.0/logo-clearml.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/partners/logo-clearml.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/social/logo-transparent.png" width="15%" height="0" alt="" />
<a href="https://bit.ly/yolov5-readme-comet">
<img src="https://github.com/ultralytics/yolov5/releases/download/v1.0/logo-comet.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/partners/logo-comet.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/social/logo-transparent.png" width="15%" height="0" alt="" />
<a href="https://bit.ly/yolov5-neuralmagic">
<img src="https://github.com/ultralytics/yolov5/releases/download/v1.0/logo-neuralmagic.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/partners/logo-neuralmagic.png" width="10%" /></a>
</div>
| Roboflow | ClearML ⭐ NEW | Comet ⭐ NEW | Neural Magic ⭐ NEW |
@ -232,7 +232,8 @@ on your experience. Thank you 🙏 to all our contributors!
<!-- SVG image from https://opencollective.com/ultralytics/contributors.svg?width=990 -->
<a href="https://github.com/ultralytics/ultralytics/graphs/contributors"><img src="https://github.com/ultralytics/yolov5/releases/download/v1.0/image-contributors-1280.png"/></a>
<a href="https://github.com/ultralytics/yolov5/graphs/contributors">
<img src="https://github.com/ultralytics/assets/raw/main/im/image-contributors.png" /></a>
## <div align="center">License</div>

@ -174,16 +174,16 @@ success = model.export(format="onnx") # 将模型导出为 ONNX 格式
<div align="center">
<a href="https://roboflow.com/?ref=ultralytics">
<img src="https://github.com/ultralytics/yolov5/releases/download/v1.0/logo-roboflow.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/partners/logo-roboflow.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/social/logo-transparent.png" width="15%" height="0" alt="" />
<a href="https://cutt.ly/yolov5-readme-clearml">
<img src="https://github.com/ultralytics/yolov5/releases/download/v1.0/logo-clearml.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/partners/logo-clearml.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/social/logo-transparent.png" width="15%" height="0" alt="" />
<a href="https://bit.ly/yolov5-readme-comet">
<img src="https://github.com/ultralytics/yolov5/releases/download/v1.0/logo-comet.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/partners/logo-comet.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/social/logo-transparent.png" width="15%" height="0" alt="" />
<a href="https://bit.ly/yolov5-neuralmagic">
<img src="https://github.com/ultralytics/yolov5/releases/download/v1.0/logo-neuralmagic.png" width="10%" /></a>
<img src="https://github.com/ultralytics/assets/raw/main/partners/logo-neuralmagic.png" width="10%" /></a>
</div>
| Roboflow | ClearML ⭐ 新 | Comet ⭐ 新 | Neural Magic ⭐ 新 |
@ -203,7 +203,8 @@ success = model.export(format="onnx") # 将模型导出为 ONNX 格式
<!-- SVG image from https://opencollective.com/ultralytics/contributors.svg?width=990 -->
<a href="https://github.com/ultralytics/ultralytics/graphs/contributors"><img src="https://github.com/ultralytics/yolov5/releases/download/v1.0/image-contributors-1280.png"/></a>
<a href="https://github.com/ultralytics/yolov5/graphs/contributors">
<img src="https://github.com/ultralytics/assets/raw/main/im/image-contributors.png" /></a>
## <div align="center">License</div>

@ -22,7 +22,7 @@ def get_version():
setup(
name="ultralytics", # name of pypi package
version=get_version(), # version of pypi package
python_requires=">=3.7,<=3.11",
python_requires=">=3.7",
license='GPL-3.0',
description='Ultralytics YOLOv8',
long_description=README,
@ -41,7 +41,9 @@ setup(
'dev':
['check-manifest', 'pytest', 'pytest-cov', 'coverage', 'mkdocs', 'mkdocstrings[python]', 'mkdocs-material']},
classifiers=[
"Development Status :: 4 - Beta",
"Intended Audience :: Developers",
"Intended Audience :: Education",
"Intended Audience :: Science/Research",
"License :: OSI Approved :: GNU General Public License v3 (GPLv3)",
"Programming Language :: Python :: 3",
@ -49,14 +51,14 @@ setup(
"Programming Language :: Python :: 3.8",
"Programming Language :: Python :: 3.9",
"Programming Language :: Python :: 3.10",
"Programming Language :: Python :: 3.11",
"Topic :: Software Development",
"Topic :: Scientific/Engineering",
"Topic :: Scientific/Engineering :: Artificial Intelligence",
"Topic :: Scientific/Engineering :: Image Recognition",
"Operating System :: POSIX :: Linux",
"Operating System :: MacOS",
"Operating System :: Microsoft :: Windows",
"Development Status :: 4 - Beta",],
"Operating System :: Microsoft :: Windows",],
keywords="machine-learning, deep-learning, vision, ML, DL, AI, YOLO, YOLOv3, YOLOv5, YOLOv8, HUB, Ultralytics",
entry_points={
'console_scripts': ['yolo = ultralytics.yolo.cfg:entrypoint', 'ultralytics = ultralytics.yolo.cfg:entrypoint']})

@ -1,6 +1,6 @@
# Ultralytics YOLO 🚀, GPL-3.0 license
__version__ = "8.0.34"
__version__ = "8.0.35"
from ultralytics.yolo.engine.model import YOLO
from ultralytics.yolo.utils.checks import check_yolo as checks

@ -13,7 +13,7 @@ import torch.nn as nn
from PIL import Image
from ultralytics.yolo.utils import LOGGER, ROOT, yaml_load
from ultralytics.yolo.utils.checks import check_requirements, check_suffix, check_version
from ultralytics.yolo.utils.checks import check_requirements, check_suffix, check_version, check_yaml
from ultralytics.yolo.utils.downloads import attempt_download_asset, is_url
from ultralytics.yolo.utils.ops import xywh2xyxy
@ -38,7 +38,7 @@ class AutoBackend(nn.Module):
weights (str): The path to the weights file. Default: 'yolov8n.pt'
device (torch.device): The device to run the model on.
dnn (bool): Use OpenCV's DNN module for inference if True, defaults to False.
data (dict): Additional data, optional
data (str), (Path): Additional data.yaml file for class names, optional
fp16 (bool): If True, use half precision. Default: False
fuse (bool): Whether to fuse the model or not. Default: True
@ -237,7 +237,7 @@ class AutoBackend(nn.Module):
# class names
if 'names' not in locals(): # names missing
names = yaml_load(data)['names'] if data else {i: f'class{i}' for i in range(999)} # assign default
names = yaml_load(check_yaml(data))['names'] if data else {i: f'class{i}' for i in range(999)} # assign
names = check_class_names(names)
self.__dict__.update(locals()) # assign all variables to self

@ -2,6 +2,7 @@
import contextlib
from copy import deepcopy
from pathlib import Path
import thop
import torch
@ -490,6 +491,14 @@ def guess_model_task(model):
with contextlib.suppress(Exception):
cfg = eval(x)
break
elif isinstance(model, (str, Path)):
model = str(model)
if '-seg' in model:
return "segment"
elif '-cls' in model:
return "classify"
else:
return "detect"
# Guess from YAML dictionary
if cfg:

@ -8,10 +8,8 @@ from pathlib import Path
from types import SimpleNamespace
from typing import Dict, List, Union
from ultralytics.yolo.utils import (DEFAULT_CFG, DEFAULT_CFG_DICT, DEFAULT_CFG_PATH, LOGGER, PREFIX, ROOT,
USER_CONFIG_DIR, IterableSimpleNamespace, __version__, colorstr, emojis, yaml_load,
yaml_print)
from ultralytics.yolo.utils.checks import check_yolo
from ultralytics.yolo.utils import (DEFAULT_CFG, DEFAULT_CFG_DICT, DEFAULT_CFG_PATH, LOGGER, ROOT, USER_CONFIG_DIR,
IterableSimpleNamespace, __version__, checks, colorstr, yaml_load, yaml_print)
CLI_HELP_MSG = \
"""
@ -83,7 +81,7 @@ def cfg2dict(cfg):
return cfg
def get_cfg(cfg: Union[str, Path, Dict, SimpleNamespace] = DEFAULT_CFG, overrides: Dict = None):
def get_cfg(cfg: Union[str, Path, Dict, SimpleNamespace] = DEFAULT_CFG_DICT, overrides: Dict = None):
"""
Load and merge configuration data from a file or dictionary.
@ -198,17 +196,23 @@ def entrypoint(debug=''):
LOGGER.info(CLI_HELP_MSG)
return
# Add tasks, modes, special, and special with dash keys, i.e. -help, --help
# Define tasks and modes
tasks = 'detect', 'segment', 'classify'
modes = 'train', 'val', 'predict', 'export'
# Define special commands
special = {
'help': lambda: LOGGER.info(CLI_HELP_MSG),
'checks': check_yolo,
'checks': checks.check_yolo,
'version': lambda: LOGGER.info(__version__),
'settings': lambda: yaml_print(USER_CONFIG_DIR / 'settings.yaml'),
'cfg': lambda: yaml_print(DEFAULT_CFG_PATH),
'copy-cfg': copy_default_cfg}
FULL_ARGS_DICT = {**DEFAULT_CFG_DICT, **{k: None for k in tasks}, **{k: None for k in modes}, **special}
full_args_dict = {**DEFAULT_CFG_DICT, **{k: None for k in tasks}, **{k: None for k in modes}, **special}
# Define common mis-uses of special commands, i.e. -h, -help, --help
special.update({k[0]: v for k, v in special.items()}) # singular
special.update({k[:-1]: v for k, v in special.items() if len(k) > 1 and k.endswith('s')}) # singular
special = {**special, **{f'-{k}': v for k, v in special.items()}, **{f'--{k}': v for k, v in special.items()}}
overrides = {} # basic overrides, i.e. imgsz=320
@ -219,7 +223,7 @@ def entrypoint(debug=''):
k, v = a.split('=', 1) # split on first '=' sign
assert v, f"missing '{k}' value"
if k == 'cfg': # custom.yaml passed
LOGGER.info(f"{PREFIX}Overriding {DEFAULT_CFG_PATH} with {v}")
LOGGER.info(f"Overriding {DEFAULT_CFG_PATH} with {v}")
overrides = {k: val for k, val in yaml_load(v).items() if k != 'cfg'}
else:
if v.lower() == 'none':
@ -233,7 +237,7 @@ def entrypoint(debug=''):
v = eval(v)
overrides[k] = v
except (NameError, SyntaxError, ValueError, AssertionError) as e:
check_cfg_mismatch(FULL_ARGS_DICT, {a: ""}, e)
check_cfg_mismatch(full_args_dict, {a: ""}, e)
elif a in tasks:
overrides['task'] = a
@ -248,7 +252,7 @@ def entrypoint(debug=''):
raise SyntaxError(f"'{colorstr('red', 'bold', a)}' is a valid YOLO argument but is missing an '=' sign "
f"to set its value, i.e. try '{a}={DEFAULT_CFG_DICT[a]}'\n{CLI_HELP_MSG}")
else:
check_cfg_mismatch(FULL_ARGS_DICT, {a: ""})
check_cfg_mismatch(full_args_dict, {a: ""})
# Defaults
task2model = dict(detect='yolov8n.pt', segment='yolov8n-seg.pt', classify='yolov8n-cls.pt')
@ -261,9 +265,9 @@ def entrypoint(debug=''):
LOGGER.warning(f"WARNING ⚠️ 'mode' is missing. Valid modes are {modes}. Using default 'mode={mode}'.")
elif mode not in modes:
if mode != 'checks':
raise ValueError(emojis(f"ERROR ❌ Invalid 'mode={mode}'. Valid modes are {modes}."))
raise ValueError(f"Invalid 'mode={mode}'. Valid modes are {modes}.")
LOGGER.warning("WARNING ⚠️ 'yolo mode=checks' is deprecated. Use 'yolo checks' instead.")
check_yolo()
checks.check_yolo()
return
# Model
@ -304,7 +308,7 @@ def entrypoint(debug=''):
def copy_default_cfg():
new_file = Path.cwd() / DEFAULT_CFG_PATH.name.replace('.yaml', '_copy.yaml')
shutil.copy2(DEFAULT_CFG_PATH, new_file)
LOGGER.info(f"{PREFIX}{DEFAULT_CFG_PATH} copied to {new_file}\n"
LOGGER.info(f"{DEFAULT_CFG_PATH} copied to {new_file}\n"
f"Example YOLO command with this new custom cfg:\n yolo cfg='{new_file}' imgsz=320 batch=8")

@ -161,8 +161,6 @@ class Exporter:
# Checks
model.names = check_class_names(model.names)
# if self.args.batch == model.args['batch_size']: # user has not modified training batch_size
self.args.batch = 1
self.imgsz = check_imgsz(self.args.imgsz, stride=model.stride, min_dim=2) # check image size
if model.task == 'classify':
self.args.nms = self.args.agnostic_nms = False

@ -57,15 +57,17 @@ class YOLO:
self.overrides = {} # overrides for trainer object
# Load or create new YOLO model
load_methods = {'.pt': self._load, '.yaml': self._new}
suffix = Path(model).suffix
if not suffix and Path(model).stem in GITHUB_ASSET_STEMS:
model, suffix = Path(model).with_suffix('.pt'), '.pt' # add suffix, i.e. yolov8n -> yolov8n.pt
if suffix in load_methods:
{'.pt': self._load, '.yaml': self._new}[suffix](model)
try:
if suffix == '.yaml':
self._new(model)
else:
raise NotImplementedError(f"'{suffix}' models not supported. Try a *.pt and *.yaml model, "
"i.e. model='yolov8n.pt' or model='yolov8n.yaml'")
self._load(model)
except Exception as e:
raise NotImplementedError(f"Unable to load model='{model}'. "
f"As an example try model='yolov8n.pt' or model='yolov8n.yaml'") from e
def __call__(self, source=None, stream=False, **kwargs):
return self.predict(source, stream, **kwargs)
@ -78,13 +80,11 @@ class YOLO:
cfg (str): model configuration file
verbose (bool): display model info on load
"""
cfg = check_yaml(cfg) # check YAML
cfg_dict = yaml_load(cfg, append_filename=True) # model dict
self.cfg = check_yaml(cfg) # check YAML
cfg_dict = yaml_load(self.cfg, append_filename=True) # model dict
self.task = guess_model_task(cfg_dict)
self.ModelClass, self.TrainerClass, self.ValidatorClass, self.PredictorClass = \
self._assign_ops_from_task(self.task)
self.ModelClass, self.TrainerClass, self.ValidatorClass, self.PredictorClass = self._assign_ops_from_task()
self.model = self.ModelClass(cfg_dict, verbose=verbose) # initialize
self.cfg = cfg
def _load(self, weights: str):
"""
@ -93,13 +93,17 @@ class YOLO:
Args:
weights (str): model checkpoint to be loaded
"""
suffix = Path(weights).suffix
if suffix == '.pt':
self.model, self.ckpt = attempt_load_one_weight(weights)
self.ckpt_path = weights
self.task = self.model.args["task"]
self.overrides = self.model.args
self._reset_ckpt_args(self.overrides)
self.ModelClass, self.TrainerClass, self.ValidatorClass, self.PredictorClass = \
self._assign_ops_from_task(self.task)
else:
self.model, self.ckpt = weights, None
self.task = guess_model_task(weights)
self.ckpt_path = weights
self.ModelClass, self.TrainerClass, self.ValidatorClass, self.PredictorClass = self._assign_ops_from_task()
def reset(self):
"""
@ -166,7 +170,7 @@ class YOLO:
args = get_cfg(cfg=DEFAULT_CFG, overrides=overrides)
args.data = data or args.data
args.task = self.task
if args.imgsz == DEFAULT_CFG.imgsz:
if args.imgsz == DEFAULT_CFG.imgsz and not isinstance(self.model, (str, Path)):
args.imgsz = self.model.args['imgsz'] # use trained imgsz unless custom value is passed
args.imgsz = check_imgsz(args.imgsz, max_dim=1)
@ -189,7 +193,8 @@ class YOLO:
args.task = self.task
if args.imgsz == DEFAULT_CFG.imgsz:
args.imgsz = self.model.args['imgsz'] # use trained imgsz unless custom value is passed
if args.batch == DEFAULT_CFG.batch:
args.batch = 1 # default to 1 if not modified
exporter = Exporter(overrides=args)
exporter(model=self.model)
@ -231,8 +236,8 @@ class YOLO:
"""
self.model.to(device)
def _assign_ops_from_task(self, task):
model_class, train_lit, val_lit, pred_lit = MODEL_MAP[task]
def _assign_ops_from_task(self):
model_class, train_lit, val_lit, pred_lit = MODEL_MAP[self.task]
# warning: eval is unsafe. Use with caution
trainer_class = eval(train_lit.replace("TYPE", f"{self.type}"))
validator_class = eval(val_lit.replace("TYPE", f"{self.type}"))

@ -146,7 +146,7 @@ class BasePredictor:
(self.save_dir / 'labels' if self.args.save_txt else self.save_dir).mkdir(parents=True, exist_ok=True)
# warmup model
if not self.done_warmup:
self.model.warmup(imgsz=(1 if self.model.pt or self.model.triton else self.bs, 3, *self.imgsz))
self.model.warmup(imgsz=(1 if self.model.pt or self.model.triton else self.dataset.bs, 3, *self.imgsz))
self.done_warmup = True
self.seen, self.windows, self.dt, self.batch = 0, [], (ops.Profile(), ops.Profile(), ops.Profile()), None
@ -218,7 +218,7 @@ class BasePredictor:
cv2.namedWindow(str(p), cv2.WINDOW_NORMAL | cv2.WINDOW_KEEPRATIO) # allow window resize (Linux)
cv2.resizeWindow(str(p), im0.shape[1], im0.shape[0])
cv2.imshow(str(p), im0)
cv2.waitKey(1) # 1 millisecond
cv2.waitKey(500 if self.batch[4].startswith('image') else 1) # 1 millisecond
def save_preds(self, vid_cap, idx, save_path):
im0 = self.annotator.result()

@ -95,7 +95,7 @@ class BaseValidator:
assert model is not None, "Either trainer or model is needed for validation"
self.device = select_device(self.args.device, self.args.batch)
self.args.half &= self.device.type != 'cpu'
model = AutoBackend(model, device=self.device, dnn=self.args.dnn, fp16=self.args.half)
model = AutoBackend(model, device=self.device, dnn=self.args.dnn, data=self.args.data, fp16=self.args.half)
self.model = model
stride, pt, jit, engine = model.stride, model.pt, model.jit, model.engine
imgsz = check_imgsz(self.args.imgsz, stride=stride)

@ -138,7 +138,7 @@ def non_max_suppression(
multi_label=False,
labels=(),
max_det=300,
nm=0, # number of masks
nc=0, # number of classes (optional)
):
"""
Perform non-maximum suppression (NMS) on a set of boxes, with support for masks and multiple labels per box.
@ -159,7 +159,7 @@ def non_max_suppression(
list contains the apriori labels for a given image. The list should be in the format
output by a dataloader, with each label being a tuple of (class_index, x1, y1, x2, y2).
max_det (int): The maximum number of boxes to keep after NMS.
nm (int): The number of masks output by the model.
nc (int): (optional) The number of classes output by the model. Any indices after this will be considered masks.
Returns:
(List[torch.Tensor]): A list of length batch_size, where each element is a tensor of
@ -178,7 +178,8 @@ def non_max_suppression(
if mps: # MPS not fully supported yet, convert tensors to CPU before NMS
prediction = prediction.cpu()
bs = prediction.shape[0] # batch size
nc = prediction.shape[1] - nm - 4 # number of classes
nc = nc or (prediction.shape[1] - 4) # number of classes
nm = prediction.shape[1] - nc - 4
mi = 4 + nc # mask start index
xc = prediction[:, 4:mi].amax(1) > conf_thres # candidates

@ -40,13 +40,12 @@ class DetectionValidator(BaseValidator):
return batch
def init_metrics(self, model):
head = model.model[-1] if self.training else model.model.model[-1]
val = self.data.get(self.args.split, '') # validation path
self.is_coco = isinstance(val, str) and val.endswith(f'coco{os.sep}val2017.txt') # is COCO dataset
self.class_map = ops.coco80_to_coco91_class() if self.is_coco else list(range(1000))
self.args.save_json |= self.is_coco and not self.training # run on final val if training COCO
self.nc = head.nc
self.names = model.names
self.nc = len(model.names)
self.metrics.names = self.names
self.metrics.plot = self.args.plots
self.confusion_matrix = ConfusionMatrix(nc=self.nc)

@ -17,10 +17,10 @@ class SegmentationPredictor(DetectionPredictor):
self.args.iou,
agnostic=self.args.agnostic_nms,
max_det=self.args.max_det,
nm=32,
nc=len(self.model.names),
classes=self.args.classes)
results = []
proto = preds[1][-1]
proto = preds[1][-1] if len(preds[1]) == 3 else preds[1] # second output is len 3 if pt, but only 1 if exported
for i, pred in enumerate(p):
shape = orig_img[i].shape if isinstance(orig_img, list) else orig_img.shape
if not len(pred):

@ -28,14 +28,12 @@ class SegmentationValidator(DetectionValidator):
return batch
def init_metrics(self, model):
head = model.model[-1] if self.training else model.model.model[-1]
val = self.data.get(self.args.split, '') # validation path
self.is_coco = isinstance(val, str) and val.endswith(f'coco{os.sep}val2017.txt') # is COCO dataset
self.class_map = ops.coco80_to_coco91_class() if self.is_coco else list(range(1000))
self.args.save_json |= self.is_coco and not self.training # run on final val if training COCO
self.nc = head.nc
self.nm = head.nm if hasattr(head, "nm") else 32
self.names = model.names
self.nc = len(model.names)
self.metrics.names = self.names
self.metrics.plot = self.args.plots
self.confusion_matrix = ConfusionMatrix(nc=self.nc)
@ -61,8 +59,9 @@ class SegmentationValidator(DetectionValidator):
multi_label=True,
agnostic=self.args.single_cls,
max_det=self.args.max_det,
nm=self.nm)
return p, preds[1][-1]
nc=self.nc)
proto = preds[1][-1] if len(preds[1]) == 3 else preds[1] # second output is len 3 if pt, but only 1 if exported
return p, proto
def update_metrics(self, preds, batch):
# Metrics

Loading…
Cancel
Save