|
|
|
|
|
|
trainer_params = dummy_config |
|
|
|
trainer = PPOTrainer("test", 0, trainer_params, True, False, 0, "0") |
|
|
|
trainer.seed = 1 |
|
|
|
policy = trainer.create_policy("test", mock_specs) |
|
|
|
policy.create_tf_graph() |
|
|
|
policy = trainer.create_policy("test", mock_specs, create_graph=True) |
|
|
|
to_load_policy = trainer.create_policy("test", mock_specs) |
|
|
|
to_load_policy.create_tf_graph() |
|
|
|
to_load_policy.init_load_weights() |
|
|
|
to_load_policy = trainer.create_policy("test", mock_specs, create_graph=True) |
|
|
|
|
|
|
|
weights = policy.get_weights() |
|
|
|
load_weights = to_load_policy.get_weights() |
|
|
|