Do I need to change max_labels?

[minkichunm]

Let's say I'm using 200 objects.

I changed a class information in voc_classes.py or coco_classes.py

In yolo_base.py and yolo_voc_s.py

    with wait_for_the_master():
        dataset = COCODataset(
            data_dir=self.data_dir,
            json_file=self.train_ann,
            img_size=self.input_size,
            preproc=TrainTransform(
                max_labels=50,
                flip_prob=self.flip_prob,
                hsv_prob=self.hsv_prob),
            cache=cache_img,
        )

    dataset = MosaicDetection(
        dataset,
        mosaic=not no_aug,
        img_size=self.input_size,
        preproc=TrainTransform(
            max_labels=120,
            flip_prob=self.flip_prob,
            hsv_prob=self.hsv_prob),
        degrees=self.degrees,
        translate=self.translate,
        mosaic_scale=self.mosaic_scale,
        mixup_scale=self.mixup_scale,
        shear=self.shear,
        enable_mixup=self.enable_mixup,
        mosaic_prob=self.mosaic_prob,
        mixup_prob=self.mixup_prob,
    )

should I need to change also max_labels in COCODataset or MosaicDetection?

[Superyanzhuang]

I also encountered this problem, I reduced amx_labels=50, but return an error

[FateScript]

It depends on the max annoation number of a single image in your dataset. If it's greater than config in YOLOX, plz set the value.