Hi, how to go around this error? ```Metaflow 2.12....
# ask-metaflow
h
Hi, how to go around this error?
Copy code
Metaflow 2.12.8 executing ModelFlow for user:sln
Validating your flow...
    The graph looks good!
2024-09-12 23:03:52.663 Bootstrapping virtual environment(s) ...
    Pip ran into an error while setting up environment:
    Could not find a binary distribution for pyod>=1.1.3 
    for the platform linux-64
    
    Note that @pypi does not currently support source distributions
Packages:
Copy code
@pypi_base(
    packages={
        "joblib": "1.2.0",
        "matplotlib": "3.5.1",
        "numpy": "1.24.0",
        "pandas": "1.5.3",
        "pillow": "10.1.0",
        "psycopg2-binary": "2.9.9",
        "requests": "2.31.0",
        "seaborn": "0.11.2",
        "torch": "2.1.2",
        "transformers": "4.41.2",
        "xgboost": "2.0.3",
        "pydantic": "2.5.3",
        "urllib3": "1.26.18",
        "pydantic-settings": "2.1.0",
        "jsonschema": "4.21.1",
        "pycaret": "3.3.2",
        "sqlalchemy": "2.0.28",
        "nltk": "3.8.1",
        "spacy": "3.7.4",
        "sentence-transformers": "3.0.1",
        "alembic": "1.13.1",
        "pgvector": "",
        "git+<https://github.com/openai/CLIP.git>": "",
        "fastapi": "",
    },
)
class ModelFlow(FlowSpec):
      ......
I need to use pypi as conda doesnt have pycaret version 3.3.2
1
d
pyod
doesn’t have a wheel. You can try the extension which supports this: https://docs.metaflow.org/scaling/dependencies/libraries#bleeding-edge-versions-of-the-decorators
h
Thanks @dry-beach-38304
Is there a config for
--max-workers
which can be added to metaflow config (generally resides in
~/.metaflowconfig/config.json
)? Want to set it as a default for all the flows. FYI - Running tasks remotely(k8s)
d
You can set an env var: METAFLOW_RUN_MAX_WORKERS if you are using the run command.
h
Thanks for the help @dry-beach-38304
no prob 1
d
did you solve your first dependency issues btw (conda/pypi)?
h
Yup. Installed metaflow netflix extension.
👍 1