rolv-arild / rocket-learn Goto Github PK

Traceback (most recent call last):
File "D:\rocket-learn\examples\default\learner.py", line 129, in
alg.run(iterations_per_save=100, save_dir="models")
File "D:\rocket-learn\rocket_learn2\ppo.py", line 188, in run
self.rollout_generator.update_parameters(self.agent.actor)
File "D:\rocket-learn\rocket_learn2\rollout_generator\redis\redis_rollout_generator.py", line 324, in update_parameters
self.redis.hset(CONTRIBUTORS, mapping=tot_contributors)
File "C:\Users\lukas\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.9_qbz5n2kfra8p0\LocalCache\local-packages\Python39\site-packages\redis\commands\core.py", line 4841, in hset
return self.execute_command("HSET", name, *items)
File "C:\Users\lukas\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.9_qbz5n2kfra8p0\LocalCache\local-packages\Python39\site-packages\redis\client.py", line 1238, in execute_command
return conn.retry.call_with_retry(
File "C:\Users\lukas\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.9_qbz5n2kfra8p0\LocalCache\local-packages\Python39\site-packages\redis\retry.py", line 46, in call_with_retry
return do()
File "C:\Users\lukas\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.9_qbz5n2kfra8p0\LocalCache\local-packages\Python39\site-packages\redis\client.py", line 1239, in
lambda: self._send_command_parse_response(
File "C:\Users\lukas\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.9_qbz5n2kfra8p0\LocalCache\local-packages\Python39\site-packages\redis\client.py", line 1215, in _send_command_parse_response
return self.parse_response(conn, command_name, **options)
File "C:\Users\lukas\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.9_qbz5n2kfra8p0\LocalCache\local-packages\Python39\site-packages\redis\client.py", line 1254, in parse_response
response = connection.read_response()
File "C:\Users\lukas\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.9_qbz5n2kfra8p0\LocalCache\local-packages\Python39\site-packages\redis\connection.py", line 839, in read_response
raise response
redis.exceptions.ResponseError: wrong number of arguments for 'hset' command

EVALUATION GAME
╭─────────────────────────────┬───────────┬──────────────────────────┬───────────╮
│ Blue │ rating │ Orange │ rating │
├─────────────────────────────┼───────────┼──────────────────────────┼───────────┤
│ Bot-v0-deterministic │ 0.00±2.00 │Bot-v0-stochastic │ 0.00±2.00 │
╰─────────────────────────────┴───────────┴──────────────────────────┴───────────╯
0 steps [00:00, ? steps/s]Traceback (most recent call last):
File "D:\rocket-learn\examples\default\worker.py", line 66, in
RedisRolloutWorker(r, "MrBean_BOT", match,
File "D:\rocket-learn\rocket_learn2\rollout_generator\redis\redis_rollout_worker.py", line 362, in run
result = rocket_learn2.utils.generate_episode.generate_episode(self.env, agents, evaluate=True,
File "D:\rocket-learn\rocket_learn2\utils\generate_episode.py", line 31, in generate_episode
game_condition = GameCondition(tick_skip=env._match._tick_skip, # noqa
TypeError: init() got an unexpected keyword argument 'seconds_per_goal_forfeit'

I think this only occurs on new bots when there are no old versions yet. I think it just needs to check that an old version exists before trying to call the rand. I can work on it eventually, just wanted to make an issue so we don't lose track.

C:\Users\user\Code\Kaiyotech\ABAD\venv\lib\site-packages\rocket_learn\rollout_generator\redis_rollout_generator.py:490: RuntimeWarning: invalid value encountered in true_divide
probs /= probs.sum()
Traceback (most recent call last):
File "C:\Users\user\AppData\Local\Programs\Python\Python39\lib\runpy.py", line 197, in _run_module_as_main
return _run_code(code, main_globals, None,
File "C:\Users\user\AppData\Local\Programs\Python\Python39\lib\runpy.py", line 87, in _run_code
exec(code, run_globals)
File "C:\Users\user\Code\Kaiyotech\ABAD\training\worker.py", line 122, in
RedisRolloutWorker(r,
File "C:\Users\user\Code\Kaiyotech\ABAD\venv\lib\site-packages\rocket_learn\rollout_generator\redis_rollout_generator.py", line 573, in run
versions = self._get_opponent_indices(n_new, n_old, pretrained_choice)
File "C:\Users\user\Code\Kaiyotech\ABAD\venv\lib\site-packages\rocket_learn\rollout_generator\redis_rollout_generator.py", line 492, in _get_opponent_indices
old_versions = np.random.choice(len(probs), size=n_old, p=probs, replace=n_new > 0).tolist()
File "mtrand.pyx", line 935, in numpy.random.mtrand.RandomState.choice
ValueError: probabilities contain NaN

Warning: Gym version v0.24.0 has a number of critical issues with gym.make such that the reset and step functions are called before returning the environment. It is recommend to downgrading to v0.23.1 or upgrading to v0.25.1
C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\win32\lib\pywintypes.py:2: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
import imp, sys, os
WARNING: The Scoreboard object overwrites the inverted ball ang.vel. to include scoreboard, make sure you're not using that and instead inverting on your own. Call it in your obs builder's pre-step method to use.
Started worker 1498d56a-1f7b-44a0-b770-6d965ef7d0f9 on host 127.0.0.1 under name test
Launching Rocket League, make sure bakkesmod is running.
Launched Epic version
Executing injector...
Scanning processes
Found 2 processes
Done
Generating rollout with versions: ['-29 (0.00±2.00)', '-29 (0.00±2.00)', '-29 (0.00±2.00)', '-29 (0.00±2.00)', '-29 (0.00±2.00)', '-29 (0.00±2.00)']
Index 2 is out of range for self.replay_setters
Problem Detected. Killing Worker...
Traceback (most recent call last):
File "worker.py", line 156, in
main()
File "worker.py", line 149, in main
worker.run()
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\rocket_learn\rollout_generator\redis\redis_rollout_worker.py", line 346, in run
Rocket League (64-bit, DX11, Cooked) (1654354) is now hidden
rollouts, result = rocket_learn.utils.generate_episode.generate_episode(self.env, agents,
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\rocket_learn\utils\generate_episode.py", line 70, in generate_episode
dist = policy.get_action_distribution(obs)
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\rocket_learn\agent\discrete_policy.py", line 42, in get_action_distribution
logits = th.Tensor(self(obs))
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\rocket_learn\agent\discrete_policy.py", line 19, in forward
logits = self.net(obs)
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\torch\nn\modules\container.py", line 217, in forward
input = module(input)
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\torch\nn\modules\linear.py", line 114, in forward
return F.linear(input, self.weight, self.bias)
TypeError: linear(): argument 'input' (position 1) must be Tensor, not tuple
Press any key to continue . . .

Hello,
currently im training an bot using this repo and I've asked myself some questions to this time. Is learning with multiple instances more efficent than just one and how can you do multiple instances (just setting up torch.set_num_threads(1))? Also is it more efficient to train it against an real bot for example yours or should it play against itself?

Thanks for answering and have a good day!
OGMatrix

pygame 2.3.0 (SDL 2.24.2, Python 3.8.6)
Hello from the pygame community. https://www.pygame.org/contribute.html
Problem Detected. Killing Worker...
Traceback (most recent call last):
File "worker.py", line 138, in
main()
File "worker.py", line 126, in main
worker = make_worker(ip, name, password,
File "worker.py", line 77, in make_worker
replay_arrays = _unserialize(r.get("replay-arrays"))
File "C:\Users\test\AppData\Local\necto3\venv\lib\site-packages\rocket_learn\rollout_generator\redis_rollout_generator.py", line 66, in _unserialize
return msgpack.unpackb(zlib.decompress(obj))
TypeError: a bytes-like object is required, not 'NoneType'

Warning: Gym version v0.24.0 has a number of critical issues with gym.make such that the reset and step functions are called before returning the environment. It is recommend to downgrading to v0.23.1 or upgrading to v0.25.1
C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\win32\lib\pywintypes.py:2: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
import imp, sys, os
WARNING: The Scoreboard object overwrites the inverted ball ang.vel. to include scoreboard, make sure you're not using that and instead inverting on your own. Call it in your obs builder's pre-step method to use.
Started worker d77cdf26-908b-4fcb-b391-f38e1e830765 on host 127.0.0.1 under name test
Launching Rocket League, make sure bakkesmod is running.
Launched Epic version
Executing injector...
Scanning processes
Found 2 processes
Done
Generating rollout with versions: ['-29 (0.00±2.00)', '-29 (0.00±2.00)']
Index 0 is out of range for self.replay_setters
Problem Detected. Killing Worker...
Traceback (most recent call last):
File "worker.py", line 156, in
main()
File "worker.py", line 149, in main
worker.run()
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\rocket_learn\rollout_generator\redis\redis_rollout_worker.py", line 346, in run
rollouts, result = rocket_learn.utils.generate_episode.generate_episode(self.env, agents,
Rocket League (64-bit, DX11, Cooked) (987476) is now hidden
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\rocket_learn\utils\generate_episode.py", line 70, in generate_episode
dist = policy.get_action_distribution(obs)
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\rocket_learn\agent\discrete_policy.py", line 42, in get_action_distribution
logits = th.Tensor(self(obs))
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\rocket_learn\agent\discrete_policy.py", line 19, in forward
logits = self.net(obs)
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\torch\nn\modules\container.py", line 217, in forward
input = module(input)
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "C:\Users\maciej\AppData\Local\necto2\venv\lib\site-packages\torch\nn\modules\linear.py", line 114, in forward
return F.linear(input, self.weight, self.bias)
TypeError: linear(): argument 'input' (position 1) must be Tensor, not tuple