vladfi1/slippi-ai

Imitation learning from slippi replays.

slippi-ai's Issues

Unable to evaluate model due to nonexistent User/Config directory in Dolphin

Hey! Thanks for this awesome project and for making it open source and so easy to use. I managed to train the fox ditto model on google colab (and happy to upload the experiment files to allow others to have a pre-trained model). However, when I try to evaluate the model on OSX I get:

python eval_self.py with saved_model_path=~/Downloads/experiments/2021-10-31_dccfa16f1a3767f6/saved_model/ dolphin_path=/Applications/Dolphin.app iso_path=isos/ssbm.iso

2021-10-31 16:53:33.641164: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
WARNING - eval - No observers have been added to this run
INFO - eval - Running command 'main'
INFO - eval - Started
ERROR - eval - Failed after 0:00:00!
Traceback (most recent calls WITHOUT Sacred internals):
  File "eval_self.py", line 27, in main
    blocking_input=True,
  File "/Users/nick/.local/lib/python3.6/site-packages/melee/console.py", line 98, in __init__
    _copytree_safe(self._get_dolphin_home_path(), temp_dir)
  File "/Users/nick/.local/lib/python3.6/site-packages/melee/console.py", line 51, in _copytree_safe
    shutil.copytree(src, dst, ignore=_ignore_fifos)
  File "/Users/nick/anaconda3/envs/hci_ssbm/lib/python3.6/shutil.py", line 315, in copytree
    names = os.listdir(src)
FileNotFoundError: [Errno 2] No such file or directory: '/Applications/Dolphin.app/Contents/Resources/User/'

I've tried creating a dummy directory manually, but then run into:

python eval_cpu.py with saved_model_path=~/Downloads/experiments/2021-10-31_dccfa16f1a3767f6/saved_model/ dolphin_path=/Applications/Dolphin.app iso_path=isos/ssbm.iso

2021-10-31 19:28:11.733256: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
WARNING - eval - No observers have been added to this run
INFO - eval - Running command 'main'
INFO - eval - Started
ERROR - eval - Failed after 0:00:00!
Traceback (most recent calls WITHOUT Sacred internals):
  File "eval_cpu.py", line 28, in main
    blocking_input=True,
  File "/Users/nick/.local/lib/python3.6/site-packages/melee/console.py", line 137, in __init__
    raise InvalidDolphinPath(self._get_dolphin_config_path())
melee.console.InvalidDolphinPath: /Applications/Dolphin.app/Contents/Resources/User/Config/

Is /Applications/Dolphin.app the right path? By the way I'm using dolphin 5.0-15260. I'm not sure if that matters.