pipeline.config 파일의 설정에 대한 페이지는 없나?
있는거 뜯어보니 상위 엘리먼트(?)는 아래와 같이 5개로 나눠진다.
이름만 보면 직관적이긴 한데 model / training / evaluation 세가지 그리고 training, evaluation에 대한 읽기 설정인 듯.
model {}
train_config {}
train_input_reader {}
eval_config {}
eval_input_reader {}
google model garden 에서 받아서 파일에서 분석을 해보니 아래와 같은 종류가 나온다.
'cosine' 'darknet' 'dilated_resnet' 'embedded_ssd_mobilenet_v1' 'exponential' 'faster_rcnn_inception_resnet_v2' 'faster_rcnn_inception_resnet_v2_keras' 'faster_rcnn_inception_v2' 'faster_rcnn_nas' 'faster_rcnn_resnet101' 'faster_rcnn_resnet101_keras' 'faster_rcnn_resnet152' 'faster_rcnn_resnet152_keras' 'faster_rcnn_resnet50' 'faster_rcnn_resnet50_fpn_keras' 'faster_rcnn_resnet50_keras' 'identity' 'linear' 'lstm_mobilenet_v1' 'lstm_mobilenet_v1_fpn' 'lstm_ssd_interleaved_mobilenet_v2' 'lstm_ssd_mobilenet_v1' 'mobilenet' 'polynomial' 'resnet' 'sgd' 'spinenet' 'ssd_efficientnet-b0_bifpn_keras' 'ssd_efficientnet-b1_bifpn_keras' 'ssd_efficientnet-b2_bifpn_keras' 'ssd_efficientnet-b3_bifpn_keras' 'ssd_efficientnet-b4_bifpn_keras' 'ssd_efficientnet-b5_bifpn_keras' 'ssd_efficientnet-b6_bifpn_keras' 'ssd_inception_v2' 'ssd_inception_v3' 'ssd_mobiledet_cpu' 'ssd_mobiledet_dsp' 'ssd_mobiledet_edgetpu' 'ssd_mobiledet_gpu' 'ssd_mobilenet_edgetpu' 'ssd_mobilenet_v1' 'ssd_mobilenet_v1_fpn' 'ssd_mobilenet_v1_fpn_keras' 'ssd_mobilenet_v1_ppn' 'ssd_mobilenet_v2' 'ssd_mobilenet_v2_fpn' 'ssd_mobilenet_v2_fpn_keras' 'ssd_mobilenet_v2_keras' 'ssd_mobilenet_v2_mnasfpn' 'ssd_mobilenet_v3_large' 'ssd_mobilenet_v3_small' 'ssd_resnet101_v1_fpn' 'ssd_resnet101_v1_fpn_keras' 'ssd_resnet152_v1_fpn_keras' 'ssd_resnet50_v1_fpn' 'ssd_resnet50_v1_fpn_keras' 'stepwise' |
아래는 research / object_detection 아래만 검색한 내용
'embedded_ssd_mobilenet_v1' 'faster_rcnn_inception_resnet_v2' 'faster_rcnn_inception_resnet_v2_keras' 'faster_rcnn_inception_v2' 'faster_rcnn_nas' 'faster_rcnn_resnet101' 'faster_rcnn_resnet101_keras' 'faster_rcnn_resnet152' 'faster_rcnn_resnet152_keras' 'faster_rcnn_resnet50' 'faster_rcnn_resnet50_fpn_keras' 'faster_rcnn_resnet50_keras' 'ssd_efficientnet-b0_bifpn_keras' 'ssd_efficientnet-b1_bifpn_keras' 'ssd_efficientnet-b2_bifpn_keras' 'ssd_efficientnet-b3_bifpn_keras' 'ssd_efficientnet-b4_bifpn_keras' 'ssd_efficientnet-b5_bifpn_keras' 'ssd_efficientnet-b6_bifpn_keras' 'ssd_inception_v2' 'ssd_inception_v3' 'ssd_mobiledet_cpu' 'ssd_mobiledet_dsp' 'ssd_mobiledet_edgetpu' 'ssd_mobiledet_gpu' 'ssd_mobilenet_edgetpu' 'ssd_mobilenet_v1' 'ssd_mobilenet_v1_fpn' 'ssd_mobilenet_v1_fpn_keras' 'ssd_mobilenet_v1_ppn' 'ssd_mobilenet_v2' 'ssd_mobilenet_v2_fpn' 'ssd_mobilenet_v2_fpn_keras' 'ssd_mobilenet_v2_keras' 'ssd_mobilenet_v2_mnasfpn' 'ssd_mobilenet_v3_large' 'ssd_mobilenet_v3_small' 'ssd_resnet101_v1_fpn' 'ssd_resnet101_v1_fpn_keras' 'ssd_resnet152_v1_fpn_keras' 'ssd_resnet50_v1_fpn' 'ssd_resnet50_v1_fpn_keras' |
'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글
ssd_mobilenet_v2 on tf1, tf2 (0) | 2021.02.12 |
---|---|
fpn - Feature Pyramid Network (0) | 2021.02.11 |
tensorflow pipeline.conf (0) | 2021.02.08 |
tf object detection COCO (0) | 2021.02.05 |
tensorflow lite SELECT_TF_OPS (0) | 2021.02.02 |