一、注册自己的数据集
使用detectron2训练自己的数据集，第一步要注册自己的数据集。
- 首先保证自己的数据集标注是coco格式，就可以使用
load_coco_json加载自己的数据集并转化为detectron2的专有数据格式。
- 使用
DatasetCatalog.register注册训练集和测试集。
- 使用
MetadataCatalog.get注册训练集和测试集的标注元数据
要纠结代码细节，具体注册数据集的逻辑思路参见下面几个函数，细节自己细品。
data/datasets/buildin.py/register_all_coco(root="datasets")
data/datasets/buildin_meta.py/_get_builtin_metadata(dataset_name)
data/datasets/buildin_meta.py/_get_coco_instances_meta()
data/datasets/register_coco.py/register_coco_instances(name, metadata, json_file, image_root)
data/datasets/coco.py/load_coco_json(json_file, image_root, dataset_name=None, extra_annotation_keys=None)
直接上代码
from detectron2.data import DatasetCatalog, MetadataCatalog
from detectron2.data.datasets.coco import load_coco_json
DatasetCatalog.register("dataset_train", lambda: load_coco_json(TRAIN_JSON, TRAIN_PATH, "dataset_train"))
MetadataCatalog.get("dataset_train").set(thing_classes=["pos", "neg"],
json_file=TRAIN_JSON,
image_root=TRAIN_PATH)
DatasetCatalog.register("dataset_val", lambda: load_coco_json(VAL_JSON, VAL_PATH, "dataset_val"))
MetadataCatalog.get("dataset_val").set(thing_classes=["pos", "neg"],
json_file=VAL_JSON,
image_root=VAL_PATH)
二、训练自己的数据集
- 在官方的
tools/train_net.py上加上注册数据集部分
- 继承
DefaultTrainer父类，定义Trainer，重写build_evaluator类方法，使得Trainer具有评估功能。
class Trainer(DefaultTrainer):
@classmethod
def build_evaluator(cls, cfg, dataset_name, output_folder=None):
if output_folder is None:
output_folder = os.path.join(cfg.OUTPUT_DIR, "inference")
return COCOEvaluator(dataset_name, cfg, distributed=False, output_dir=output_folder)
detectron2的训练器框架，具体细节可通过F12查看DefaultTrainer的父类SimpleTrainer的父类TrainerBase。是一个从抽像到具象的过程，通过重写父类方法累加功能。
我的训练脚本（包含注册数据集）
import os
import cv2
import logging
from collections import OrderedDict
import detectron2.utils.comm as comm
from detectron2.utils.visualizer import Visualizer
from detectron2.checkpoint import DetectionCheckpointer
from detectron2.config import get_cfg
from detectron2.data import DatasetCatalog, MetadataCatalog
from detectron2.data.datasets.coco import load_coco_json
from detectron2.engine import DefaultTrainer, default_argument_parser, default_setup, launch
from detectron2.evaluation import COCOEvaluator, verify_results
from detectron2.modeling import GeneralizedRCNNWithTTA
# 数据集路径
DATASET_ROOT = '/home/Documents/Datasets'
ANN_ROOT = os.path.join(DATASET_ROOT, 'annotations')
TRAIN_PATH = os.path.join(DATASET_ROOT, 'train')
VAL_PATH = os.path.join(DATASET_ROOT, 'val')
TRAIN_JSON = os.path.join(ANN_ROOT, 'train.json')
VAL_JSON = os.path.join(ANN_ROOT, 'val.json')
# 数据集类别元数据
DATASET_CATEGORIES = [
{"name": "pos", "id": 1, "isthing": 1, "color": [220, 20, 60]},
{"name": "neg", "id": 2, "isthing": 1, "color": [219, 142, 185]},
]
# 数据集的子集
PREDEFINED_SPLITS_DATASET = {
"train_2019": (TRAIN_PATH, TRAIN_JSON),
"val_2019": (VAL_PATH, VAL_JSON),
}
def register_dataset():
"""
purpose: register all splits of dataset with PREDEFINED_SPLITS_DATASET
"""
for key, (image_root, json_file) in PREDEFINED_SPLITS_DATASET.items():
register_dataset_instances(name=key,
metadate=get_dataset_instances_meta(),
json_file=json_file,
image_root=image_root)
def get_dataset_instances_meta():
"""
purpose: get metadata of dataset from DATASET_CATEGORIES
return: dict[metadata]
"""
thing_ids = [k["id"] for k in DATASET_CATEGORIES if k["isthing"] == 1]
thing_colors = [k["color"] for k in DATASET_CATEGORIES if k["isthing"] == 1]
# assert len(thing_ids) == 2, len(thing_ids)
thing_dataset_id_to_contiguous_id = {k: i for i, k in enumerate(thing_ids)}
thing_classes = [k["name"] for k in DATASET_CATEGORIES if k["isthing"] == 1]
ret = {
"thing_dataset_id_to_contiguous_id": thing_dataset_id_to_contiguous_id,
"thing_classes": thing_classes,
"thing_colors": thing_colors,
}
return ret
def register_dataset_instances(name, metadate, json_file, image_root):
"""
purpose: register dataset to DatasetCatalog,
register metadata to MetadataCatalog and set attribute
"""
DatasetCatalog.register(name, lambda: load_coco_json(json_file, image_root, name))
MetadataCatalog.get(name).set(json_file=json_file,
image_root=image_root,
evaluator_type="coco",
**metadate)
# 注册数据集和元数据
def plain_register_dataset():
DatasetCatalog.register("train_2019", lambda: load_coco_json(TRAIN_JSON, TRAIN_PATH, "train_2019"))
MetadataCatalog.get("train_2019").set(thing_classes=["pos", "neg"],
json_file=TRAIN_JSON,
image_root=TRAIN_PATH)
DatasetCatalog.register("val_2019", lambda: load_coco_json(VAL_JSON, VAL_PATH, "val_2019"))
MetadataCatalog.get("val_2019").set(thing_classes=["pos", "neg"],
json_file=VAL_JSON,
image_root=VAL_PATH)
# 查看数据集标注
def checkout_dataset_annotation(name="val_2019"):
dataset_dicts = load_coco_json(TRAIN_JSON, TRAIN_PATH, name)
for d in dataset_dicts:
img = cv2.imread(d["file_name"])
visualizer = Visualizer(img[:, :, ::-1], metadata=MetadataCatalog.get(name), scale=1.5)
vis = visualizer.draw_dataset_dict(d)
cv2.imshow('show', vis.get_image()[:, :, ::-1])
cv2.waitKey(0)
class Trainer(DefaultTrainer):
@classmethod
def build_evaluator(cls, cfg, dataset_name, output_folder=None):
if output_folder is None:
output_folder = os.path.join(cfg.OUTPUT_DIR, "inference")
return COCOEvaluator(dataset_name, cfg, distributed=False, output_dir=output_folder)
@classmethod
def test_with_TTA(cls, cfg, model):
logger = logging.getLogger("detectron2.trainer")
# In the end of training, run an evaluation with TTA
# Only support some R-CNN models.
logger.info("Running inference with test-time augmentation ...")
model = GeneralizedRCNNWithTTA(cfg, model)
evaluators = [
cls.build_evaluator(
cfg, name, output_folder=os.path.join(cfg.OUTPUT_DIR, "inference_TTA")
)
for name in cfg.DATASETS.TEST
]
res = cls.test(cfg, model, evaluators)
res = OrderedDict({k + "_TTA": v for k, v in res.items()})
return res
def setup(args):
"""
Create configs and perform basic setups.
"""
cfg = get_cfg() # 拷贝default config副本
args.config_file = "configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_1x.yaml"
cfg.merge_from_file(args.config_file) # 从config file 覆盖配置
cfg.merge_from_list(args.opts) # 从CLI参数 覆盖配置
# 更改配置参数
cfg.DATASETS.TRAIN = ("train_2019",)
cfg.DATASETS.TEST = ("val_2019",)
cfg.DATALOADER.NUM_WORKERS = 0 # 单线程
cfg.INPUT.MAX_SIZE_TRAIN = 400
cfg.INPUT.MAX_SIZE_TEST = 400
cfg.INPUT.MIN_SIZE_TRAIN = (160,)
cfg.INPUT.MIN_SIZE_TEST = 160
cfg.MODEL.ROI_HEADS.NUM_CLASSES = 2 # 类别数
cfg.MODEL.WEIGHTS = "/homeDocuments/pretrainedModel/Detectron2/R-50.pkl" # 预训练模型权重
cfg.SOLVER.IMS_PER_BATCH = 2 # batch_size=2; iters_in_one_epoch = dataset_imgs/batch_size
ITERS_IN_ONE_EPOCH = int(1434 / cfg.SOLVER.IMS_PER_BATCH)
cfg.SOLVER.MAX_ITER = (ITERS_IN_ONE_EPOCH * 12) - 1 # 12 epochs
cfg.SOLVER.BASE_LR = 0.002
cfg.SOLVER.MOMENTUM = 0.9
cfg.SOLVER.WEIGHT_DECAY = 0.0001
cfg.SOLVER.WEIGHT_DECAY_NORM = 0.0
cfg.SOLVER.GAMMA = 0.1
cfg.SOLVER.STEPS = (30000,)
cfg.SOLVER.WARMUP_FACTOR = 1.0 / 1000
cfg.SOLVER.WARMUP_ITERS = 1000
cfg.SOLVER.WARMUP_METHOD = "linear"
cfg.SOLVER.CHECKPOINT_PERIOD = ITERS_IN_ONE_EPOCH - 1
cfg.freeze()
default_setup(cfg, args)
return cfg
def main(args):
cfg = setup(args)
print(cfg)
# 注册数据集
register_dataset()
if args.eval_only:
model = Trainer.build_model(cfg)
DetectionCheckpointer(model, save_dir=cfg.OUTPUT_DIR).resume_or_load(
cfg.MODEL.WEIGHTS, resume=args.resume
)
res = Trainer.test(cfg, model)
if comm.is_main_process():
verify_results(cfg, res)
return res
trainer = Trainer(cfg)
trainer.resume_or_load(resume=args.resume)
return trainer.train()
if __name__ == "__main__":
args = default_argument_parser().parse_args()
print("Command Line Args:", args)
launch(
main,
args.num_gpus,
num_machines=args.num_machines,
machine_rank=args.machine_rank,
dist_url=args.dist_url,
args=(args,),
)
三、推断数据
推断数据的脚本是根据官方的demo.py改的。
在推断之前，也是要注册自己的数据集。
我的推断脚本
import argparse
import glob
import multiprocessing as mp
import os
import time
import cv2
import tqdm
from detectron2.config import get_cfg
from detectron2.data import MetadataCatalog, DatasetCatalog
from detectron2.data.datasets import load_coco_json
from detectron2.data.detection_utils import read_image
from detectron2.utils.logger import setup_logger
from detectron2.utils.visualizer import ColorMode
from demo.predictor import VisualizationDemo
# constants
WINDOW_NAME = "detections"
# inference
INPUT_IMG_PATH = '/home/Documents/mask_THzDatasets/val'
OUTPUT_IMG_PATH = '/home/Documents/out'
# 数据集路径
DATASET_ROOT = '/home/Documents/Datasets'
ANN_ROOT = os.path.join(DATASET_ROOT, 'annotations')
TRAIN_PATH = os.path.join(DATASET_ROOT, 'train')
VAL_PATH = os.path.join(DATASET_ROOT, 'val')
TRAIN_JSON = os.path.join(ANN_ROOT, 'train.json')
VAL_JSON = os.path.join(ANN_ROOT, 'val.json')
# 数据集类别元数据
DATASET_CATEGORIES = [
{"name": "pos", "id": 1, "isthing": 1, "color": [220, 20, 60]},
{"name": "neg", "id": 2, "isthing": 1, "color": [219, 142, 185]},
]
# 数据集的子集
PREDEFINED_SPLITS_DATASET = {
"train_2019": (TRAIN_PATH, TRAIN_JSON),
"val_2019": (VAL_PATH, VAL_JSON),
}
def register_dataset():
"""
purpose: register all splits of dataset with PREDEFINED_SPLITS_DATASET
"""
for key, (image_root, json_file) in PREDEFINED_SPLITS_DATASET.items():
register_dataset_instances(name=key,
metadate=get_dataset_instances_meta(),
json_file=json_file,
image_root=image_root)
def get_dataset_instances_meta():
"""
purpose: get metadata of dataset from DATASET_CATEGORIES
return: dict[metadata]
"""
thing_ids = [k["id"] for k in DATASET_CATEGORIES if k["isthing"] == 1]
thing_colors = [k["color"] for k in DATASET_CATEGORIES if k["isthing"] == 1]
# assert len(thing_ids) == 2, len(thing_ids)
thing_dataset_id_to_contiguous_id = {k: i for i, k in enumerate(thing_ids)}
thing_classes = [k["name"] for k in DATASET_CATEGORIES if k["isthing"] == 1]
ret = {
"thing_dataset_id_to_contiguous_id": thing_dataset_id_to_contiguous_id,
"thing_classes": thing_classes,
"thing_colors": thing_colors,
}
return ret
def register_dataset_instances(name, metadate, json_file, image_root):
"""
purpose: register dataset to DatasetCatalog,
register metadata to MetadataCatalog and set attribute
"""
DatasetCatalog.register(name, lambda: load_coco_json(json_file, image_root, name))
MetadataCatalog.get(name).set(json_file=json_file,
image_root=image_root,
evaluator_type="coco",
**metadate)
# 注册数据集和元数据
def plain_register_dataset():
DatasetCatalog.register("train_2019", lambda: load_coco_json(TRAIN_JSON, TRAIN_PATH, "train_2019"))
MetadataCatalog.get("train_2019").set(thing_classes=["pos", "neg"],
json_file=TRAIN_JSON,
image_root=TRAIN_PATH)
DatasetCatalog.register("val_2019", lambda: load_coco_json(VAL_JSON, VAL_PATH, "val_2019"))
MetadataCatalog.get("val_2019").set(thing_classes=["pos", "neg"],
json_file=VAL_JSON,
image_root=VAL_PATH)
def setup_cfg(args):
# load config from file and command-line arguments
cfg = get_cfg()
args.config_file = "configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_1x.yaml"
cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
cfg.DATASETS.TRAIN = ("train_2019",)
cfg.DATASETS.TEST = ("val_2019",)
cfg.DATALOADER.NUM_WORKERS = 0
cfg.INPUT.MAX_SIZE_TRAIN = 400
cfg.INPUT.MAX_SIZE_TEST = 400
cfg.INPUT.MIN_SIZE_TRAIN = (160,)
cfg.INPUT.MIN_SIZE_TEST = 160
cfg.MODEL.ROI_HEADS.NUM_CLASSES = 2 # 类别数
# cfg.MODEL.WEIGHTS = "/home/Documents/pretrainedModel/Detectron2/R-50.pkl" # 预训练模型权重
cfg.MODEL.WEIGHTS = '/home/anaconda3/envs/pytorch13/detectron2/output/model_final.pth' # 最终权重
cfg.SOLVER.IMS_PER_BATCH = 2 # batch_size=2; iteration = 1434/batch_size = 717 iters in one epoch
ITERS_IN_ONE_EPOCH = int(1434 / cfg.SOLVER.IMS_PER_BATCH)
cfg.SOLVER.MAX_ITER = (ITERS_IN_ONE_EPOCH * 12) - 1 # 12 epochs
cfg.SOLVER.BASE_LR = 0.002
cfg.SOLVER.MOMENTUM = 0.9
cfg.SOLVER.WEIGHT_DECAY = 0.0001
cfg.SOLVER.WEIGHT_DECAY_NORM = 0.0
cfg.SOLVER.GAMMA = 0.1
cfg.SOLVER.STEPS = (30000,)
cfg.SOLVER.WARMUP_FACTOR = 1.0 / 1000
cfg.SOLVER.WARMUP_ITERS = 1000
cfg.SOLVER.WARMUP_METHOD = "linear"
cfg.SOLVER.CHECKPOINT_PERIOD = ITERS_IN_ONE_EPOCH - 1
cfg.freeze()
return cfg
def get_parser():
parser = argparse.ArgumentParser(description="Detectron2 Demo")
parser.add_argument(
"--config-file",
default="configs/quick_schedules/e2e_mask_rcnn_R_50_FPN_inference_acc_test.yaml",
metavar="FILE",
help="path to config file",
)
parser.add_argument("--input", nargs="+", help="A list of space separated input images")
parser.add_argument(
"--output",
help="A file or directory to save output visualizations. "
"If not given, will show output in an OpenCV window.",
)
parser.add_argument(
"--confidence-threshold",
type=float,
default=0.5,
help="Minimum score for instance predictions to be shown",
)
parser.add_argument(
"--opts",
help="Modify config options using the command-line 'KEY VALUE' pairs",
default=[],
nargs=argparse.REMAINDER,
)
return parser
if __name__ == "__main__":
mp.set_start_method("spawn", force=True)
args = get_parser().parse_args()
logger = setup_logger()
logger.info("Arguments: " + str(args))
cfg = setup_cfg(args)
# 注册数据集
register_dataset()
demo = VisualizationDemo(cfg, instance_mode=ColorMode.SEGMENTATION)
# for path in tqdm.tqdm(args.input, disable=not args.output):
for imgfile in os.listdir(INPUT_IMG_PATH):
# use PIL, to be consistent with evaluation
img_fullName = os.path.join(INPUT_IMG_PATH, imgfile)
img = read_image(img_fullName, format="BGR")
start_time = time.time()
predictions, visualized_output = demo.run_on_image(img)
logger.info(
"{}: detected {} instances in {:.2f}s".format(
imgfile, len(predictions["instances"]), time.time() - start_time
)
)
if args.output:
if os.path.isdir(args.output):
assert os.path.isdir(args.output), args.output
out_filename = os.path.join(args.output, os.path.basename(imgfile))
else:
assert len(args.input) == 1, "Please specify a directory with args.output"
out_filename = args.output
visualized_output.save(out_filename)
else:
cv2.namedWindow(WINDOW_NAME, cv2.WINDOW_NORMAL)
cv2.imshow(WINDOW_NAME, visualized_output.get_image()[:, :, ::-1])
if cv2.waitKey(0) == 27:
break # esc to quit
