Skip to content

Instantly share code, notes, and snippets.

[net]
# Testing
# batch=1
# subdivisions=1
# Training
batch=6
subdivisions=1
width=608
height=608
channels=3
/content/images/190596972_AMI022888__.jpg
/content/images/190580794_AMI273089__2.jpg
/content/images/190581042_AMI106796__.jpg
/content/images/190581081_AMI063674__.jpg
/content/images/190583060_AMI271026__.jpg
/content/images/190587883_AMI091414__.jpg
/content/images/190499166_AMI182030__.jpg
/content/images/190578129_AMI279791__1.jpg
/content/images/190578129_AMI279791__2.jpg
/content/images/190594877_AMI280017__.jpg
/content/images/190561621_AMI083265__.jpg
/content/images/190565558_AMI059039__.jpg
/content/images/190587038_AMI005897__1.jpg
/content/images/190587038_AMI005897__2.jpg
/content/images/190399663_AMI216505__2.jpg
/content/images/190408911_AMI061137__1.jpg
/content/images/190450134_AMI077993__.jpg
/content/images/190439933_AMI275885__.jpg
/content/images/190400375_AMI109769__.jpg
/content/images/190420713_AMI275404__.jpg
classes= 15
train = /content/train.txt
valid = /content/val.txt
names = /content/voc.names
backup = backup
单证名称Document Type
件数No. of Packages
毛重Gross Weight
体积Cube
包装 Packages
目的港 Final Destination
装运港 Place of Loading
发货人 Shipper
收货人 Consignee
通知人 Notify Party
[net]
# Testing
# batch=1
# subdivisions=1
# Training
batch=64
subdivisions=16
width=608
height=608
channels=3
shipper
consignee
email
classes= 3
train = /home/botree/study/amass2/YOLO/model_training/merge_26/train.txt
valid = /home/botree/study/amass2/YOLO/model_training/merge_26/val.txt
names = /home/botree/study/amass2/YOLO/model_training/voc.names
backup = backupyol
import cv2
import numpy as np
import pytesseract
# Load your trained model.
net = cv2.dnn.readNet("/home/botree/study/amass2/YOLO/models/yolov3-data_final_26.weights","/home/botree/study/amass2/YOLO/model_training/yolov3-data.cfg")
print(net)
classes = ["shipper","consignee","email"]
from detectron2.data import DatasetCatalog, MetadataCatalog
for type in ['train','val']:
DatasetCatalog.register("amass_images_"+type, lambda type=type: get_amass_dicts(type))
MetadataCatalog.get("amass_images_" + type).set(thing_classes=["shipper", "consignee", "email"])
amass_metadata = MetadataCatalog.get("amass_images_train")