Data format

Open trucleduc opened this issue 9 years ago • 1 comments

Does anyone know how data is stored in both Moving MNIST and UFC101 datasets? I want to try on my own data.

Thanks,

Jan 23 '17 19:01 trucleduc

Hi, I am reframing my question, I got the structure of your data in .h5 file as Dataset 'test' Size: 784x10000 MaxSize: 784x10000 Datatype: H5T_IEEE_F32LE (single) ChunkSize: [] Filters: none FillValue: 0.000000 Dataset 'test_labels' Size: 10000 MaxSize: 10000 Datatype: H5T_STD_I64LE (int64) ChunkSize: [] Filters: none FillValue: 0 Dataset 'train' Size: 784x50000 MaxSize: 784x50000 Datatype: H5T_IEEE_F32LE (single) ChunkSize: [] Filters: none FillValue: 0.000000 Dataset 'train_full' Size: 784x60000 MaxSize: 784x60000 Datatype: H5T_IEEE_F32LE (single) ChunkSize: [] Filters: none FillValue: 0.000000 Dataset 'train_full_labels' Size: 60000 MaxSize: 60000 Datatype: H5T_STD_I64LE (int64) ChunkSize: [] Filters: none FillValue: 0 Dataset 'train_labels' Size: 50000 MaxSize: 50000 Datatype: H5T_STD_I64LE (int64) ChunkSize: [] Filters: none FillValue: 0 Dataset 'validation' Size: 784x10000 MaxSize: 784x10000 Datatype: H5T_IEEE_F32LE (single) ChunkSize: [] Filters: none FillValue: 0.000000 Dataset 'validation_labels' Size: 10000 MaxSize: 10000 Datatype: H5T_STD_I64LE (int64) ChunkSize: [] Filters: none FillValue: 0

I wanted to know how can I get a similar structure for my data? Could you give me some hint, how I could progress, Also, what do you mean by train, train_full, train_full_labels, Thanks :)

Mar 08 '18 17:03 Nd-sole