I am trying to do the evaluation of a trained SSD_Mobilenetv2 320x320 fpnlite on tensorflow. I ran training and evaluation parallel in two different colabs account. But I am always getting the metrics values as -1(evaluation result) after each checkpoint. The lose is also increasing after evaluating each checkpoint.
Below is the command used to run the evaluation:
!python /model_main_tf2.py \
--pipeline_config_path /pipeline.config \
--model_dir /model_ \
--checkpoint_dir /model_
