-
用自己数据集训练Mask_RCNN出现以下错误,该如何解决哇：
lnvalidArgumentError (see above for traceback): indices[1919] = 37079 is not in [0, 36828)
[[Node: ROI/Gather_2 = Gather[Tindices=DT_INT32, Tparams=DT_FLOAT, validate_indices=true, _device="/job:localhost/replica:0/task:0/device:CPU:0"](ROI/strided_slice_6, ROI/strided_slice_7)]]
