Not able to train DetectionModelTrainer in TinyYolo

[ruvelrp]

from imageai.Detection.Custom import DetectionModelTrainer

trainer = DetectionModelTrainer()
trainer.setModelTypeAsTinyYOLOv3()
trainer.setDataDirectory(data_directory="xyz_images")
trainer.setTrainConfig(object_names_array=["xyz"], batch_size=4, num_experiments=50,train_from_pretrained_model="yolo-tiny.h5")
trainer.trainModel()

I am getting exception

---

AttributeError Traceback (most recent call last)
in ()
2
3 trainer = DetectionModelTrainer()
----> 4 trainer.setModelTypeAsTinyYOLOv3()
5 trainer.setDataDirectory(data_directory="xyz_images")
6 #trainer.setTrainConfig(object_names_array=["xyz"], batch_size=4, num_experiments=50)

AttributeError: 'DetectionModelTrainer' object has no attribute 'setModelTypeAsTinyYOLOv3'

Please look into it and please help me I am in need :/

[OlafenwaMoses]

Kindly note that for training custom detection models, only YOLOv3 is supported for now.

[ruvelrp]

Can we expect TinyYolo Training anytime soon?

[OlafenwaMoses]

We can't promise that yet. Timeline for implementation has not been decided.

[rola93]

shouldn't this be closed?

[OlafenwaMoses]

Thanks @rola93

[joshberry]

@OlafenwaMoses @rola93 Should this issue have really been closed? I see that tiny yolo has been implemented for the detector, but I don't see any indication in the source code that it's been implemented for training, which is what @ruvelrp reported.

I too am unable to figure out how to train a custom model using tiny yolo. If it's there and I'm missing it, can you please point me to an example or section of the source code where it's implemented.

Thanks! ❤️ the project so far...

[rola93]

As far as I understand, the issue is solved: the model can't be loaded because it's not implemented

[nguyenanhtuan1008]

I have same issue like this. Does anyone implement for Yolo v3 Tiny model?

[Wheest]

I have an open pull request which tries to implement this. I've got most of the way, but have a couple of bugs I haven't been able to bash yet. Thoughts or comments appreciated.