How to evaluate the trained models?

[Deepak Nayak]

Hi,

I have trained your code for arround 50k steps, using the script "python train.py --logtostderr --pipeline_config_path=data/configs/motion_rcnn_vkitti_cam.config --train_dir=output/train/motion_rcnn_vkitti_cam --gpu 0".

I wanted to know how to evaluate the trained model. I ran another script "python eval.py --logtostderr --pipeline_config_path=data/configs/motion_rcnn_vkitti_cam.config --checkpoint_dir=output/train/motion_rcnn_vkitti_cam --eval_dir=output/eval/motion_rcnn_vkitti_cam", which generated few events file in the folder.

Kindly help, in getting to know how can i evaluate.

[Simon Meister]

Hi, you should use TensorBoard to open the logs.

[Deepak Nayak]

Hi,

How to do the inference?

[chowkamlee81]

Do we have any sample code to do inference on multiple images? I hope inference code is not the same as in tesnorflow object detection api...
Kindly provide inputs...

[Simon Meister]

Inference is done like in the tensorflow object detection API, sorry.