Closed zh9586 closed 3 years ago
Hi @huihuicainiao, Thanks for your interest. The eval.py is supposed to monitor the checkpoint_dir and evaluate new checkpoints if the training produces one. You can kill it if you are not training at the same time.
Thank you very much for your reply.
Thank you for release your great work. I have try run python3 eval.py configs/car_auto_T3_train_eval_config The function falls into an endless loop.Continuous output “'Checkpoint ./checkpoints/car_auto_T3_train/model-1400001 has been evaluated already, wait for 60.000000 seconds'”. What do you suggest?