Chengfang Lu
e72bc061c5
Major changes: - New frontend (platform/web/): Vite + React 18 + TypeScript + Tailwind - 4-module navigation: 数据送标 / 模型管理 / 车队管理 / 系统管理 - Data catalog with charts (DMS/ADAS/Lane 3-tab view) - Quality review workflow (标注质检): Good/Fine/Bad scoring with auto-advance - Audit enhancements: batch operations, rejection categories, Feishu notifications - Operation audit log (操作日志) - World model simulation studio (仿真工坊) - Dataset version management with snapshots and diff - ADAS 7-class dataset integration (138K images organized + compressed) - User management with Feishu integration and pagination - CRUD/search/filter on all pages, card layout redesign - PIL-optimized image overlay rendering - Auto-snapshot on build, in_review workflow stage - Removed embedded algorithm code (now in workspace)
64 lines
2.2 KiB
Python
64 lines
2.2 KiB
Python
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
|
|
"""
|
|
Interface for Baidu's RT-DETR, a Vision Transformer-based real-time object detector.
|
|
|
|
RT-DETR offers real-time performance and high accuracy, excelling in accelerated backends like CUDA with TensorRT.
|
|
It features an efficient hybrid encoder and IoU-aware query selection for enhanced detection accuracy.
|
|
|
|
References:
|
|
https://arxiv.org/pdf/2304.08069.pdf
|
|
"""
|
|
|
|
from ultralytics.engine.model import Model
|
|
from ultralytics.nn.tasks import RTDETRDetectionModel
|
|
from ultralytics.utils.torch_utils import TORCH_1_11
|
|
|
|
from .predict import RTDETRPredictor
|
|
from .train import RTDETRTrainer
|
|
from .val import RTDETRValidator
|
|
|
|
|
|
class RTDETR(Model):
|
|
"""Interface for Baidu's RT-DETR model, a Vision Transformer-based real-time object detector.
|
|
|
|
This model provides real-time performance with high accuracy. It supports efficient hybrid encoding, IoU-aware query
|
|
selection, and adaptable inference speed.
|
|
|
|
Attributes:
|
|
model (str): Path to the pre-trained model.
|
|
|
|
Methods:
|
|
task_map: Return a task map for RT-DETR, associating tasks with corresponding Ultralytics classes.
|
|
|
|
Examples:
|
|
Initialize RT-DETR with a pre-trained model
|
|
>>> from ultralytics import RTDETR
|
|
>>> model = RTDETR("rtdetr-l.pt")
|
|
>>> results = model("image.jpg")
|
|
"""
|
|
|
|
def __init__(self, model: str = "rtdetr-l.pt") -> None:
|
|
"""Initialize the RT-DETR model with the given pre-trained model file.
|
|
|
|
Args:
|
|
model (str): Path to the pre-trained model. Supports .pt, .yaml, and .yml formats.
|
|
"""
|
|
assert TORCH_1_11, "RTDETR requires torch>=1.11"
|
|
super().__init__(model=model, task="detect")
|
|
|
|
@property
|
|
def task_map(self) -> dict:
|
|
"""Return a task map for RT-DETR, associating tasks with corresponding Ultralytics classes.
|
|
|
|
Returns:
|
|
(dict): A dictionary mapping task names to Ultralytics task classes for the RT-DETR model.
|
|
"""
|
|
return {
|
|
"detect": {
|
|
"predictor": RTDETRPredictor,
|
|
"validator": RTDETRValidator,
|
|
"trainer": RTDETRTrainer,
|
|
"model": RTDETRDetectionModel,
|
|
}
|
|
}
|