search

evilsocket / pwnagotchi

(⌐■_■) - Deep Reinforcement Learning instrumenting bettercap for WiFi pwning.
https://pwnagotchi.ai/
Other
7.58k stars 1.14k forks source link

[BUG] Update breaks the brain #731

Closed hmax42 closed 4 years ago

hmax42 commented 4 years ago

Describe the bug updating from 1.3.0 to 1.4.1 leads to: [2019-12-15 20:10:51,983] [INFO] [ai] loading /root/brain.nn ... [2019-12-15 20:10:52,259] [ERROR] error while starting AI Traceback (most recent call last): File "/usr/local/lib/python3.7/dist-packages/pwnagotchi/ai/__init__.py", line 51, in load a2c.load(config['path'], env) File "/usr/local/lib/python3.7/dist-packages/stable_baselines/common/base_class.py", line 700, in load model.set_env(env) File "/usr/local/lib/python3.7/dist-packages/stable_baselines/common/base_class.py", line 96, in set_env "Error: the environment passed must have at least the same observation space as the model was trained on." AssertionError: Error: the environment passed must have at least the same observation space as the model was trained on.

To Reproduce this behaviour was observed via:

Environment

Expected behavior no error message and loading of the AI after updating

hmax42 commented 4 years ago

now i can even find the error even in a fresh 1.3.0 with restored backup.

is there a way to fix the error?

Evg33 commented 4 years ago

rm /root/brain.nn

hmax42 commented 4 years ago

tried that. didn't get recreated. hmm, maybe i did not wait long enough... will try

Evg33 commented 4 years ago

@hmax42 wait 10 minutes after boot....

hmax42 commented 4 years ago

so

  1. attached it in manual mode
  2. stop the service
  3. rename brain.nn to brain.nn.old, but keep the json
  4. force start auto mode
  5. after 2 minutes my raspi a3+ started training and failed
  6. let it run for 20 mins -> no brain.nn
  7. update to 1.4.1 via git/pip3
  8. same error
[2019-12-17 17:00:44,876] [INFO] entering auto mode ...
...
[2019-12-17 17:02:45,787] [INFO] [ai] model created:
[2019-12-17 17:02:45,788] [INFO]       gamma: 0.99
[2019-12-17 17:02:45,789] [INFO]       n_steps: 1
[2019-12-17 17:02:45,790] [INFO]       vf_coef: 0.25
[2019-12-17 17:02:45,791] [INFO]       ent_coef: 0.01
[2019-12-17 17:02:45,792] [INFO]       max_grad_norm: 0.5
[2019-12-17 17:02:45,793] [INFO]       learning_rate: 0.001
[2019-12-17 17:02:45,794] [INFO]       alpha: 0.99
[2019-12-17 17:02:45,795] [INFO]       epsilon: 1e-05
[2019-12-17 17:02:45,796] [INFO]       verbose: 1
[2019-12-17 17:02:45,797] [INFO]       lr_schedule: constant
[2019-12-17 17:02:45,798] [INFO]       tensorboard_log: None
[2019-12-17 17:02:45,837] [INFO] [ai] learning for 50 epochs ...
[2019-12-17 17:02:45,852] [ERROR] [ai] error while training
Traceback (most recent call last):
  File "/usr/local/lib/python3.7/dist-packages/pwnagotchi/ai/train.py", line 177, in _ai_worker
    self._model.learn(total_timesteps=epochs_per_episode, callback=self.on_ai_training_step)
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/a2c/a2c.py", line 231, in learn
    runner = A2CRunner(self.env, self, n_steps=self.n_steps, gamma=self.gamma)
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/a2c/a2c.py", line 310, in __init__
    super(A2CRunner, self).__init__(env=env, model=model, n_steps=n_steps)
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/common/runners.py", line 19, in __init__
    self.obs[:] = env.reset()
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/common/vec_env/dummy_vec_env.py", line 52, in reset
    self._save_obs(env_idx, obs)
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/common/vec_env/dummy_vec_env.py", line 71, in _save_obs
    self.buf_obs[key][env_idx] = obs
...running now for over 20 mins
[2019-12-17 18:23:52,911] [WARNING] 20 epochs with no activity -> bored
...updated to version 1.4.1
...reboot
[2019-12-17 18:27:16,023] [INFO] grid plugin loaded.
...
[2019-12-17 18:27:29,002] [INFO] entering manual mode ...
...
[2019-12-17 18:28:04,176] [INFO] entering auto mode ...
...
[2019-12-17 18:30:02,440] [INFO] [ai] model created:
[2019-12-17 18:30:02,449] [INFO]       gamma: 0.99
[2019-12-17 18:30:02,452] [INFO]       n_steps: 1
[2019-12-17 18:30:02,459] [INFO]       vf_coef: 0.25
[2019-12-17 18:30:02,466] [INFO]       ent_coef: 0.01
[2019-12-17 18:30:02,468] [INFO]       max_grad_norm: 0.5
[2019-12-17 18:30:02,470] [INFO]       learning_rate: 0.001
[2019-12-17 18:30:02,471] [INFO]       alpha: 0.99
[2019-12-17 18:30:02,476] [INFO]       epsilon: 1e-05
[2019-12-17 18:30:02,483] [INFO]       verbose: 1
[2019-12-17 18:30:02,489] [INFO]       lr_schedule: constant
[2019-12-17 18:30:02,507] [INFO]       tensorboard_log: None
[2019-12-17 18:30:02,650] [INFO] [ai] learning for 50 epochs ...
[2019-12-17 18:30:02,682] [ERROR] [ai] error while training
Traceback (most recent call last):
  File "/usr/local/lib/python3.7/dist-packages/pwnagotchi/ai/train.py", line 177, in _ai_worker
    self._model.learn(total_timesteps=epochs_per_episode, callback=self.on_ai_training_step)
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/a2c/a2c.py", line 231, in learn
    runner = A2CRunner(self.env, self, n_steps=self.n_steps, gamma=self.gamma)
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/a2c/a2c.py", line 310, in __init__
    super(A2CRunner, self).__init__(env=env, model=model, n_steps=n_steps)
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/common/runners.py", line 19, in __init__
    self.obs[:] = env.reset()
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/common/vec_env/dummy_vec_env.py", line 52, in reset
    self._save_obs(env_idx, obs)
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/common/vec_env/dummy_vec_env.py", line 71, in _save_obs
    self.buf_obs[key][env_idx] = obs
dadav commented 4 years ago

Could you pls try it with the latest code?

Do this to update:

cd /usr/local/src/pwnagotchi
git pull
pip3 install .
hmax42 commented 4 years ago

i already did. that was the way i updated

...updated to version 1.4.1 the "1.4.1" part is now obviously wrong. i updated it as you wrote it.

hmax42 commented 4 years ago

git pull just set the new version number today, but i got an extra line in the log maybe i overlooked it before

[2019-12-19 20:56:41,763] [ERROR] [ai] error while training
Traceback (most recent call last):
  File "/usr/local/lib/python3.7/dist-packages/pwnagotchi/ai/train.py", line 177, in _ai_worker
    self._model.learn(total_timesteps=epochs_per_episode, callback=self.on_ai_training_step)
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/a2c/a2c.py", line 231, in learn
    runner = A2CRunner(self.env, self, n_steps=self.n_steps, gamma=self.gamma)
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/a2c/a2c.py", line 310, in __init__
    super(A2CRunner, self).__init__(env=env, model=model, n_steps=n_steps)
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/common/runners.py", line 19, in __init__
    self.obs[:] = env.reset()
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/common/vec_env/dummy_vec_env.py", line 52, in reset
    self._save_obs(env_idx, obs)
  File "/usr/local/lib/python3.7/dist-packages/stable_baselines/common/vec_env/dummy_vec_env.py", line 71, in _save_obs
    self.buf_obs[key][env_idx] = obs
ValueError: could not broadcast input array from shape (428) into shape (1,503)

okay, so the numbers from the last line are identical in the last 5 stacktraces and i found this in the log

[2019-12-01 14:26:26,208] [INFO] the last session lasted 26 hours, 9 minutes, 8 seconds (4 completed epochs, trained for 0), average reward:0.23065476190476192 (min:-0.2 max:0.4369047619047619)
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@[2019-12-07 18:20:15,116] [INFO] [update] plugin loaded.
xBelladonna commented 4 years ago

I have been able to reproduce this issue on my unit, in fact I could not figure out why it was happening or trigger it to happen reliably. I became frustrated and erased my SD card, reinstalling fresh 1.4.1 image, boot in manual mode to stop all pwnagotchi-related services by systemd, then update each component by hand separately. Then upon second boot in auto mode, baby starts working properly!

I know this is not either a graceful or entirely logical solution but I thought it useful to share since I had same problem and aforementioned procedure fixed it for me.

EDIT: the reason I update components manually instead of allowing auto update to work is that for some reason auto update goes to infinite loop without lock when updating pwngrid. Maybe should log an issue for this but I believe recent release has fixed the problem, only can wait for the next update and watch as it is installed.

hmax42 commented 4 years ago

i reset the pi 3a+ now with just a fresh 1.4.3 image and the old config.yml. still the same error (same numbers)

i start to believe the pi 3 a+ is to blame. but i am pretty sure it worked with the early images.

more testing

hmax42 commented 4 years ago

success with a clean 1.0.1 image and this config.yml

main:
    plugins:
      grid:
        enabled: false
      auto-update:
        enabled: false
personality:
    deauth: false
ui:
    display:
        enabled: true
        rotation: 180
        type: 'waveshare_1'

letting it create the brain and updating to 1.4.3 breaks it again

retry: fresh 1.0.1 image updates to 1.1.1 to 1.2.0 and still works after updating to 1.4.3 tells me " the environment passed must have at least the same observation space as the model was trained on." reverting to 1.3.0 via git does not change that error

all my tests point to this: between 1.2.1 and 1.3.0 something changed looks like i'm staying on 1.2.1 for this one

ghost commented 4 years ago

-- Logs begin at Wed 2020-01-01 08:06:25 GMT. -- Jan 01 08:18:42 Keram0n pwnagotchi-launcher[749]: [2020-01-01 08:18:42,647] [WARNING] From /usr/local/lib/python3.7/dist-packages/tensorflow_estimator/python/estimator/api/_v1/estimator/init.py:10: The name tf.estimator.inputs is deprecated. Please use tf.compat.v1.estimator.inputs instead. Jan 01 08:19:37 Keram0n pwnagotchi-launcher[749]: [2020-01-01 08:19:37,878] [INFO] [epoch 4] duration=00:01:15 slept_for=00:01:00 blind=0 inactive=5 active=0 peers=0 tot_bond=0.00 avg_bond=0.00 hops=0 missed=0 deauths=0 assocs=0 handshakes=0 cpu=96% mem=50% temperature=45C reward=-0.2 Jan 01 08:20:52 Keram0n pwnagotchi-launcher[749]: [2020-01-01 08:20:52,152] [INFO] [epoch 5] duration=00:01:14 slept_for=00:01:00 blind=0 inactive=6 active=0 peers=0 tot_bond=0.00 avg_bond=0.00 hops=0 missed=0 deauths=0 assocs=0 handshakes=0 cpu=96% mem=50% temperature=45C reward=-0.2 Jan 01 08:22:08 Keram0n pwnagotchi-launcher[749]: [2020-01-01 08:22:08,245] [INFO] [epoch 6] duration=00:01:16 slept_for=00:01:00 blind=0 inactive=7 active=0 peers=0 tot_bond=0.00 avg_bond=0.00 hops=0 missed=0 deauths=0 assocs=0 handshakes=0 cpu=97% mem=50% temperature=45C reward=-0.2 Jan 01 08:23:07 Keram0n pwnagotchi-launcher[749]: [2020-01-01 08:23:07,251] [WARNING] !!! captured new handshake on channel 11, -82 dBm: 98:b8:ba:29:cf:85 () -> WIN_502536 [fc:2b:b2:3e:9a:02 ()] !!! Jan 01 08:23:10 Keram0n systemd[1]: Stopping pwnagotchi Deep Reinforcement Learning instrumenting bettercap for WiFI pwning.... Jan 01 08:23:10 Keram0n systemd[1]: pwnagotchi.service: Main process exited, code=killed, status=15/TERM Jan 01 08:23:10 Keram0n systemd[1]: pwnagotchi.service: Succeeded. Jan 01 08:23:10 Keram0n systemd[1]: Stopped pwnagotchi Deep Reinforcement Learning instrumenting bettercap for WiFI pwning.. Jan 01 08:23:10 Keram0n systemd[1]: Started pwnagotchi Deep Reinforcement Learning instrumenting bettercap for WiFI pwning.. Jan 01 08:23:23 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:23,654] [INFO] memtemp plugin loaded. Jan 01 08:23:23 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:23,666] [INFO] Session-stats plugin loaded. Jan 01 08:23:23 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:23,669] [INFO] [led] plugin loaded for /sys/class/leds/led0/brightness Jan 01 08:23:23 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:23,671] [INFO] grid plugin loaded. Jan 01 08:23:23 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:23,730] [WARNING] ui.fps is 0, the display will only update for major changes Jan 01 08:23:23 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:23,746] [INFO] initializing waveshare v2 display Jan 01 08:23:26 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:26,116] [INFO] [ai] loading /root/brain.json Jan 01 08:23:26 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:26,195] [INFO] Keram0n@ecaae949f513d80c80bc31dd9493bdccee10a43418c16d647594b418e86be3e6 (v1.4.1) Jan 01 08:23:26 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:26,211] [INFO] entering auto mode ... Jan 01 08:23:26 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:26,244] [INFO] [ai] bootstrapping dependencies ... Jan 01 08:23:26 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:26,421] [INFO] web ui available at http://0.0.0.0:8080/ Jan 01 08:23:27 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:27,521] [INFO] connecting to http://localhost:8081/api ... Jan 01 08:23:29 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:29,927] [INFO] found monitor interface: mon0 Jan 01 08:23:29 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:29,946] [INFO] supported channels: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14] Jan 01 08:23:29 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:29,980] [INFO] handshakes will be collected inside /boot/handshakes Jan 01 08:23:48 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:23:48,933] [INFO] [epoch 0] duration=00:00:22 slept_for=00:00:00 blind=0 inactive=1 active=0 peers=0 tot_bond=0.00 avg_bond=0.00 hops=0 missed=0 deauths=0 assocs=0 handshakes=0 cpu=97% mem=30% temperature=45C reward=-0.2 Jan 01 08:24:39 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:24:39,305] [WARNING] From /usr/lib/python3/dist-packages/tensorflow/init.py:98: The name tf.AUTO_REUSE is deprecated. Please use tf.compat.v1.AUTO_REUSE instead. Jan 01 08:24:39 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:24:39,320] [WARNING] From /usr/lib/python3/dist-packages/tensorflow/init.py:98: The name tf.AttrValue is deprecated. Please use tf.compat.v1.AttrValue instead. Jan 01 08:24:39 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:24:39,326] [WARNING] From /usr/lib/python3/dist-packages/tensorflow/init.py:98: The name tf.COMPILER_VERSION is deprecated. Please use tf.version.COMPILER_VERSION instead. Jan 01 08:24:39 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:24:39,332] [WARNING] From /usr/lib/python3/dist-packages/tensorflow/init.py:98: The name tf.CXX11_ABI_FLAG is deprecated. Please use tf.sysconfig.CXX11_ABI_FLAG instead. Jan 01 08:24:39 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:24:39,337] [WARNING] From /usr/lib/python3/dist-packages/tensorflow/init.py:98: The name tf.ConditionalAccumulator is deprecated. Please use tf.compat.v1.ConditionalAccumulator instead. Jan 01 08:25:15 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:25:15,097] [INFO] [ai] creating model ... Jan 01 08:25:15 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:25:15,529] [WARNING] From /usr/local/lib/python3.7/dist-packages/stable_baselines/common/policies.py:421: flatten (from tensorflow.python.layers.core) is deprecated and will be removed in a future version. Jan 01 08:25:15 Keram0n pwnagotchi-launcher[2860]: Instructions for updating: Jan 01 08:25:15 Keram0n pwnagotchi-launcher[2860]: Use keras.layers.flatten instead. Jan 01 08:25:15 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:25:15,591] [WARNING] From /usr/lib/python3/dist-packages/tensorflow_core/python/layers/core.py:333: Layer.apply (from tensorflow.python.keras.engine.base_layer) is deprecated and will be removed in a future version. Jan 01 08:25:15 Keram0n pwnagotchi-launcher[2860]: Instructions for updating: Jan 01 08:25:15 Keram0n pwnagotchi-launcher[2860]: Please use layer.__call__ method instead. Jan 01 08:25:16 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:25:16,141] [WARNING] From /usr/local/lib/python3.7/dist-packages/tensorflow_estimator/python/estimator/api/_v1/estimator/init.py:10: The name tf.estimator.inputs is deprecated. Please use tf.compat.v1.estimator.inputs instead. Jan 01 08:28:03 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:28:03,069] [INFO] [epoch 1] duration=00:04:14 slept_for=00:00:30 blind=0 inactive=2 active=0 peers=0 tot_bond=0.00 avg_bond=0.00 hops=0 missed=0 deauths=0 assocs=0 handshakes=0 cpu=98% mem=50% temperature=46C reward=-0.2 Jan 01 08:29:06 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:29:06,003] [WARNING] From /usr/local/lib/python3.7/dist-packages/stable_baselines/a2c/a2c.py:135: The name tf.summary.scalar is deprecated. Please use tf.compat.v1.summary.scalar instead. Jan 01 08:31:02 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:31:02,437] [WARNING] From /usr/lib/python3/dist-packages/tensorflow_core/python/ops/clip_ops.py:286: add_dispatch_support..wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version. Jan 01 08:31:02 Keram0n pwnagotchi-launcher[2860]: Instructions for updating: Jan 01 08:31:02 Keram0n pwnagotchi-launcher[2860]: Use tf.where in 2.0, which has the same broadcast rule as np.where Jan 01 08:31:03 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:31:03,497] [WARNING] From /usr/local/lib/python3.7/dist-packages/stable_baselines/a2c/a2c.py:159: The name tf.train.RMSPropOptimizer is deprecated. Please use tf.compat.v1.train.RMSPropOptimizer instead. Jan 01 08:31:03 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:31:03,511] [WARNING] From /usr/lib/python3/dist-packages/tensorflow_core/python/training/rmsprop.py:119: calling Ones.init (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version. Jan 01 08:31:03 Keram0n pwnagotchi-launcher[2860]: Instructions for updating: Jan 01 08:31:03 Keram0n pwnagotchi-launcher[2860]: Call initializer instance with the dtype argument instead of passing it to the constructor Jan 01 08:31:27 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:31:27,017] [WARNING] From /usr/local/lib/python3.7/dist-packages/stable_baselines/a2c/a2c.py:171: The name tf.summary.merge_all is deprecated. Please use tf.compat.v1.summary.merge_all instead. Jan 01 08:31:27 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:31:27,121] [INFO] [ai] loading /root/brain.nn ... Jan 01 08:31:27 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:31:27,475] [ERROR] error while starting AI Jan 01 08:31:27 Keram0n pwnagotchi-launcher[2860]: Traceback (most recent call last): Jan 01 08:31:27 Keram0n pwnagotchi-launcher[2860]: File "/usr/local/lib/python3.7/dist-packages/pwnagotchi/ai/init.py", line 51, in load Jan 01 08:31:27 Keram0n pwnagotchi-launcher[2860]: a2c.load(config['path'], env) Jan 01 08:31:27 Keram0n pwnagotchi-launcher[2860]: File "/usr/local/lib/python3.7/dist-packages/stable_baselines/common/base_class.py", line 700, in load Jan 01 08:31:27 Keram0n pwnagotchi-launcher[2860]: model.set_env(env) Jan 01 08:31:27 Keram0n pwnagotchi-launcher[2860]: File "/usr/local/lib/python3.7/dist-packages/stable_baselines/common/base_class.py", line 98, in set_env Jan 01 08:31:27 Keram0n pwnagotchi-launcher[2860]: "Error: the environment passed must have at least the same action space as the model was trained on." Jan 01 08:31:27 Keram0n pwnagotchi-launcher[2860]: AssertionError: Error: the environment passed must have at least the same action space as the model was trained on. Jan 01 08:31:27 Keram0n pwnagotchi-launcher[2860]: [2020-01-01 08:31:27,570] [WARNING] [ai] AI not loaded!

This issue just started on a Raspberry pi Zero W. Using v1.4.1 on Buster. My logs if this helps anyone.

ghost commented 4 years ago

This issue was fixed for me after removing brain.nn from /root

Jan 01 09:58:25 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:25,423] [INFO] [ai] model created: Jan 01 09:58:25 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:25,438] [INFO] gamma: 0.99 Jan 01 09:58:25 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:25,442] [INFO] n_steps: 1 Jan 01 09:58:25 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:25,452] [INFO] vf_coef: 0.25 Jan 01 09:58:25 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:25,456] [INFO] ent_coef: 0.01 Jan 01 09:58:25 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:25,467] [INFO] max_grad_norm: 0.5 Jan 01 09:58:25 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:25,471] [INFO] learning_rate: 0.001 Jan 01 09:58:25 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:25,482] [INFO] alpha: 0.99 Jan 01 09:58:25 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:25,485] [INFO] epsilon: 1e-05 Jan 01 09:58:25 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:25,496] [INFO] verbose: 1 Jan 01 09:58:25 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:25,501] [INFO] lr_schedule: constant Jan 01 09:58:25 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:25,536] [INFO] tensorboard_log: None Jan 01 09:58:26 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:26,099] [INFO] [ai] learning for 50 epochs ... Jan 01 09:58:54 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:58:54,620] [INFO] [epoch 18] duration=00:01:05 slept_for=00:01:00 blind=0 inactive=13 active=0 peers=0 tot_bond=0.00 avg_bond=0.00 hops=0 missed=0 deauths=0 assocs=0 handshakes=0 cpu=63% mem=60% temperature=52C reward=-0.1368421052631579 Jan 01 09:59:08 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:08,841] [INFO] [ai] setting new policy: Jan 01 09:59:09 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:09,009] [INFO] [ai] ! min_rssi: -200 -> -70 Jan 01 09:59:09 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:09,176] [INFO] [ai] ! ap_ttl: 120 -> 341 Jan 01 09:59:09 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:09,215] [INFO] [ai] ! sta_ttl: 300 -> 141 Jan 01 09:59:09 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:09,265] [INFO] [ai] ! recon_time: 30 -> 52 Jan 01 09:59:09 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:09,339] [INFO] [ai] ! max_inactive_scale: 2 -> 6 Jan 01 09:59:09 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:09,380] [INFO] [ai] ! recon_inactive_multiplier: 2 -> 3 Jan 01 09:59:09 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:09,404] [INFO] [ai] ! hop_recon_time: 10 -> 12 Jan 01 09:59:09 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:09,435] [INFO] [ai] ! min_recon_time: 5 -> 30 Jan 01 09:59:09 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:09,446] [INFO] [ai] ! max_misses_for_recon: 5 -> 7 Jan 01 09:59:09 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:09,489] [INFO] [ai] ! excited_num_epochs: 10 -> 14 Jan 01 09:59:09 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:09,520] [INFO] [ai] ! bored_num_epochs: 15 -> 7 Jan 01 09:59:09 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:09,536] [INFO] [ai] ! sad_num_epochs: 25 -> 6 Jan 01 09:59:09 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:09,570] [INFO] [ai] ! channels: [] -> [1, 3, 4, 9, 10, 13, 14] Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,075] [INFO] [ai] --- training epoch 1/50 --- Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,082] [INFO] [ai] REWARD: -0.136842 Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,105] [INFO] [ai] observation: Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,135] [INFO] aps Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,139] [INFO] CH 1: 0.3163265306119221 Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,148] [INFO] CH 2: 0.010204081632642649 Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,160] [INFO] CH 4: 0.010204081632642649 Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,177] [INFO] CH 5: 0.010204081632642649 Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,193] [INFO] CH 6: 0.15306122448963974 Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,208] [INFO] CH 8: 0.020408163265285297 Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,229] [INFO] CH 9: 0.030612244897927948 Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,233] [INFO] CH 10: 0.010204081632642649 Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,243] [INFO] CH 11: 0.43877551020363387 Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,247] [INFO] sta Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,258] [INFO] CH 6: 0.4374999999972657 Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,262] [INFO] CH 8: 0.06249999999960938 Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,275] [INFO] CH 11: 0.49999999999687506 Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,286] [INFO] peers Jan 01 09:59:10 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:10,308] [INFO] [ai] saving model to /root/brain.nn ... Jan 01 09:59:17 Keram0n pwnagotchi-launcher[6448]: [2020-01-01 09:59:17,338] [INFO] [ai] saving /root/brain.json

dadav commented 4 years ago

Derek close

qkum commented 3 years ago

I can confirm 18-3-2021 that the fix is to remove the brain.nn file -> reboot -> it recreates it when the AI is done loading. You can easily track it on the log plugin in the browser.

Just encountered this problem within 24 hours after installing pwnagotchi on my Raspberry pi 4 with the official Pwnagotchi 1/2/3 image.

Command to delete the file: sudo rm /root/brain.nn

Then reboot