Dataset about backbone

[FengMu1995]

I wonder why the backbone was trained with only one dataset every time. Why can't the backbone be trained with multiple dataset. As you upload "ResNet50@WebFace600K"

Looking forward to your reply

[nttstar]

What do you mean by multiple datasets?

[FengMu1995]

Yes, the multiple datasets are combined into one big dataset which be used to train or the datasets are trained one by one.

[jacqueline-weng]

I implemented this function in my code. Make each dataset into the form of mxnet style train.rec and train.idx. When accessing to each index, convert the global index to local index which can be applied to individual dataset. The big assumption is you do not have any overlapping in id between different datasets. It worked pretty well for me.

[jacqueline-weng]

I implemented this function in my code. Make each dataset into the form of mxnet style train.rec and train.idx. When accessing to each index, convert the global index to local index which can be applied to individual dataset. The big assumption is you do not have any overlapping in id between different datasets. It worked pretty well for me.

In this way, you don't have to rewrite training rec each time adding new data.