Base installation instructions in the avalon/README.md for Ubuntu are not correct

[dfm794]

Specifically, following these two lines under ubuntu will fail
pip install avalon-rl
python -m avalon.install_godot_binary

With this stack trace
python -m avalon.install_godot_binary
Traceback (most recent call last):
File "/home/dfm/anaconda3/envs/avalon-rst/lib/python3.9/runpy.py", line 197, in _run_module_as_main
return _run_code(code, main_globals, None,
File "/home/dfm/anaconda3/envs/avalon-rst/lib/python3.9/runpy.py", line 87, in _run_code
exec(code, run_globals)
File "/home/dfm/dev/avalon-benchmark/avalon/avalon/install_godot_binary.py", line 16, in
from avalon.datagen.godot_env.interactive_godot_process import GODOT_BINARY_PATH
File "/home/dfm/dev/avalon-benchmark/avalon/avalon/datagen/godot_env/interactive_godot_process.py", line 22, in
from avalon.common.log_utils import log_to_sentry
File "/home/dfm/dev/avalon-benchmark/avalon/avalon/common/log_utils.py", line 13, in
from avalon.contrib.utils import FILESYSTEM_ROOT
File "/home/dfm/dev/avalon-benchmark/avalon/avalon/contrib/utils.py", line 18, in
import torch
ModuleNotFoundError: No module named 'torch'

It seems that torch is not installed when installing avalon-rl.
install_godot_binary uses code in utils.py, and that code imports torch for a single function that sets the random number generator seeds.

using pip install avalon-rl[train] will install torch and then the binary installation and common.check_install also succeeds.

[bai-generally-intelligent]

Ah yes you're right, for now please treat torch as a dependency for Avalon.