Continue my Training process about marllib

[Aequatio-Space]

in run_cc.py you can see that restore_path is being linked to testing.

MARLlib/marllib/marl/algos/run_cc.py

Lines 34 to 48 in 368c617

	if exp_info['restore_path']['model_path'] == '':
	restore_config = None
	else:
	restore_config = exp_info['restore_path']
	if 'render' in exp_info['restore_path']:
	render_config = {
	"evaluation_interval": 1,
	"evaluation_num_episodes": 100,
	"evaluation_num_workers": 1,
	"evaluation_config": {
	"record_env": False,
	"render_env": True,
	}
	}

To restore training progress, perhaps try resume=True in the ray.init? Or you can also rewrite the relevant logic to circumvent this issue.