Model checkpoints saved as .hdf5 instead of .tlt since 5.0.0 update

Please provide the following information when requesting support.

• Hardware (Xavier)
• Network Type (Detectnet_v2)
• TAO 5.0.0
• Training spec file(
train_config.txt (2.7 KB)

since the TAO 5.0.0 update, model checkpoints during training are saved as .hdf5 instead of .tlt. Do I need to make any changes to my train_config.txt?

Thank you very much,

It is not needed. You can download latest TAO 5.0 notebook for detectnet_v2.TAO Toolkit Getting Started | NVIDIA NGC

1 Like

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.