Hello there,
I'm training a Custom Detection Model with two types of objects in the images to test out ImageAI.
This is the Code:
from imageai.Detection.Custom import DetectionModelTrainer` trainer = DetectionModelTrainer()trainer.setModelTypeAsYOLOv3()trainer.setDataDirectory(data_directory="/content/drive/My Drive/MelipoAI/Dataset/melipo_ai")trainer.setTrainConfig(object_names_array=["jatai"], batch_size=4, num_experiments=200,`train_from_pretrained_model="yolov3.pt")trainer.trainModel()
It starts training the model but validation values are always at 0.000000
Generating anchor boxes for training images...
thr=0.25: 1.0000 best possible recall, 7.42 anchors past thr
n=9, img_size=416, metric_all=0.467/0.833-mean/best, past_thr=0.527-mean:
pretrained weight loading failed. Defaulting to using random weight.
Pretrained YOLOv3 model loaded to initialize weights
Hello there, I'm training a Custom Detection Model with two types of objects in the images to test out ImageAI. This is the Code:
trainer.setModelTypeAsYOLOv3()
trainer.setDataDirectory(data_directory="/content/drive/My Drive/MelipoAI/Dataset/melipo_ai")
trainer.setTrainConfig(object_names_array=["jatai"], batch_size=4, num_experiments=200,
from imageai.Detection.Custom import DetectionModelTrainer
`
trainer = DetectionModelTrainer()`train_from_pretrained_model="yolov3.pt")
trainer.trainModel()
It starts training the model but validation values are always at 0.000000
Generating anchor boxes for training images... thr=0.25: 1.0000 best possible recall, 7.42 anchors past thr n=9, img_size=416, metric_all=0.467/0.833-mean/best, past_thr=0.527-mean: pretrained weight loading failed. Defaulting to using random weight.
Pretrained YOLOv3 model loaded to initialize weights
Epoch 1/200
Train: 25it [00:03, 7.41it/s] box loss-> 0.10574, object loss-> 0.57634, class loss-> 0.00000 Validation: 16it [01:14, 4.64s/it] recall: 0.000000 precision: 0.000000 mAP@0.5: 0.000000, mAP@0.5-0.95: 0.000000
Epoch 2/200
Train: 25it [00:03, 7.57it/s] box loss-> 0.10004, object loss-> 0.18873, class loss-> 0.00000 Validation: 16it [01:17, 4.82s/it] recall: 0.000000 precision: 0.000000 mAP@0.5: 0.000000, mAP@0.5-0.95: 0.000000`
...and so on. What could be the reason? How can I fix this? Thanks