Difference between pretrain_model_file VS pretrain_model_path in the spec files?

What is the difference between pretrain_model_file VS pretrain_model_path in the spec files?

Which network did you talk about?

hdf5 detectnet and also .tlt custom. I have observed that _path works with .tlt but not _file.

Do you talk about TLT detectnet_v2 network? If not, is it Faster_rcnn , SSD , or others?

detectnet_v2 network

In detectnet_v2 network, please see https://docs.nvidia.com/tlt/tlt-user-guide/text/object_detection/detectnet_v2.html#model-config
there is only “pretrained_model_file”.

Usually you can set similar to below.

model_config {
pretrained_model_file: “/workspace/tlt-experiments/detectnet_v2/pretrained_resnet18/tlt_pretrained_detectnet_v2_vresnet18/resnet18.hdf5”

You can download jupyter notebook for reference. There are spec files in each network.

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.