OfflineRL icon indicating copy to clipboard operation
OfflineRL copied to clipboard

[Solved] Something went wrong in `get_repo` & Aim up don't work well

Open Dilettante258 opened this issue 2 years ago • 0 comments

$ python examples/train_task.py --algo_name=cql --exp_name=halfcheetah --task HalfCheetah-v3 --task_data_type low --task_train_num 100
2023-11-03 at 16:35:56.112 | INFO | Use cql algorithm!
running build_ext
2023-11-03 at 16:35:56.618 | INFO | obs shape: (1000, 18)
......
......
2023-11-03 at 16:35:56.870 | INFO | Init AlgoTrainer
Something went wrong in `get_repo`. The process will continue to execute.
`get_repo`: branch name must be at least 2 characters and contain only latin letters, numbers, dash and underscore
Something went wrong in `__init__`. The process will continue to execute.
`__init__`: 'NoneType' object has no attribute 'path'
Something went wrong in `track`. The process will continue to execute.
`track`: session is closed

I tried to follow the steup requirements ("aim==2.0.27"), but some nuisance still happened.

After investigation, the error was located on line 306 of the Session class in the aim up module.

Amend the file "/home/username/anaconda3/envs/env_name/lib/python3.8/site-packages/aim/sdk/session/session.py"

if path is not None:
    repo = AimRepo(path)
    if not repo.exists():
        if not repo.init():
            raise ValueError('can not create repo `{}`'.format(path))
    repo = AimRepo(path, branch_name, commit_hash)

Replace the second line of the above code with repo=AimRepo (path, branchname, commit_hash)(or delete it), but it still don't work very well.

After about 6 hours of tortuous process😵, I finally get a hapharzcan env which works well.🥳

The following table shows some versions of libraries that I think are worth emphasizing.

Name Version
python 3.8.18
cython 0.29
setuptools 63.2.0
pysqlite3 0.5.2
pip 23.3.1
ray 1.12.0
aim 2.3.0
fire 0.3.0
glfw 2.6.2
numpy 1.20.3
protobuf 3.19.0
torch 2.1.0
scikit-learn 1.3.2
mujoco-py 1.50.1.0
swig 4.1.1
gym 0.19.0

The following table is the environment profile after I deleted some unimportant libraries for your reference. 🤝I hope it can be helpful to you!

#Name Version
_libgcc_mutex 0.1
_openmp_mutex 5.1
absl-py 2.0.0
aim 2.3.0
aimrecords 0.0.7
aimrocks 0.0.7
aiofiles 23.2.1
aiohttp 3.7.4
aiohttp-cors 0.7.0
aioredis 1.3.1
aiosignal 1.3.1
asttokens 2.4.1
async-exit-stack 1.0.1
async-generator 1.10
async-timeout 3.0.1
atari-py 0.2.6
attrdict 2.0.1
attrs 23.1.0
box2d-py 2.3.5
cython 0.29
decorator 4.4.2
deprecated 1.2.14
distlib 0.3.7
dm-tree 0.1.8
docker 6.1.3
filelock 3.13.1
fire 0.3.0
flask 1.1.2
glfw 2.6.2
google-api-core 2.12.0
google-auth 1.35.0
google-auth-oauthlib 0.4.6
googleapis-common-protos 1.61.0
gtimer 1.0.0b5
gunicorn 20.1.0
gym 0.19.0
gym-notices 0.0.8
libgcc-ng 11.2.0
libgomp 11.2.0
libstdcxx-ng 11.2.0
lockfile 0.12.2
mujoco-py 1.50.1.0
multidict 6.0.4
neorl 0.3.1
networkx 3.1
numpy 1.20.3
numpydoc 1.6.0
oauthlib 3.2.2
offlinerl 0.0.1
opencensus 0.11.3
opencv-python 4.8.1.78
openssl 3.0.11
pandas 2.0.3
parso 0.8.3
pexpect 4.8.0
pickleshare 0.7.5
pillow 10.0.1
pip 23.3.1
protobuf 3.19.0
pyasn1 0.5.0
pyasn1-modules 0.3.0
pycparser 2.21
pydantic 1.10.13
pygame 2.1.0
pyglet 2.0.9
pygments 2.16.1
pyopengl 3.1.7
pyopengl-accelerate 3.1.7
pyparsing 3.1.1
pyrser 0.2.0
pysqlite3 0.5.2
python 3.8.18
ray 1.12.0
readline 8.2
redis 4.1.4
rsa 4.9
scikit-learn 1.3.2
scipy 1.10.1
setuptools 63.2.0
sphinx 7.1.2
sqlalchemy 1.4.13
sqlite 3.41.2
stack-data 0.6.3
starlette 0.14.2
swig 4.1.1
sympy 1.12
tabulate 0.9.0
tensorboard 2.3.0
tensorboardx 2.6.2.2
termcolor 2.3.0
threadpoolctl 3.2.0
tk 8.6.12
tomli 2.0.1
torch 2.1.0
tqdm 4.66.1
urllib3 2.0.7

Dilettante258 avatar Nov 03 '23 10:11 Dilettante258