While creating the spec file for efficientnet image classification , “num_layers” value for efficeintnet is not defined in the TLTV3 user guide.
model_config {
arch: “efficientnet_b0”
n_layers:
Also getting error while training “bash: tlt command not found” . Command used for training is "tlt classification train "
Please use n_layers .
See Image Classification — Transfer Learning Toolkit 3.0 documentation
arch: “resnet”
#
for resnet → n_layers can be [10, 18, 50]
#
for vgg → n_layers can be [16, 19]
n_layers: 101
I will sync with internal team to modify the user guide.
For “bash: tlt command not found”, please see TLT Launcher — Transfer Learning Toolkit 3.0 documentation, you need to install tlt. The most important step is
pip3 install nvidia-pyindex
pip3 install nvidia-tlt