Hi, Im facing the below error when running the flow:

❯ METAFLOW_CONDA_DEPENDENCY_RESOLVER="mamba" python NeuralNetFlow.py --no-pylint --datastore=s3 --environment=conda run --with kubernetes
Metaflow 2.12.8 executing NeuralNetFlow for user:sln
Validating your flow...
    The graph looks good!
2024-08-29 22:08:35.872 Bootstrapping virtual environment(s) ...
    Micromamba ran into an error while setting up environment:
    command '/home/sln/.metaflowconfig/micromamba/bin/micromamba create --yes --quiet --dry-run --no-extra-safety-checks --repodata-ttl=86400 --retry-clean-cache --prefix=/tmp/tmpypqo9jiy/prefix --channel=defaults --channel=conda-forge requests==>=2.21.0 boto3==>=1.14.0 tensorflow==2.11.0 sentence-transformers==3.0.1 matplotlib==3.4.2 pandas==1.5.3 numpy==1.25.2 psycopg2-binary==2.9.9 transformers==4.41.2 openssl==3.1.2 libpq==12.15 python==3.9.13' returned error (1)
    package tensorflow-2.11.0-cpu_py39h4655687_0 requires tensorflow-base 2.11.0 cpu_py39_0, but none of the providers can be installed

Here's the packages added:

@conda_base(
    packages={
        "tensorflow": "2.11.0",
        "sentence-transformers": "3.0.1",
        "matplotlib": "3.4.2",
        "pandas": "1.5.3",
        "numpy": "1.25.2",
        "psycopg2-binary": "2.9.9",
        "transformers": "4.41.2",
        "openssl": "3.1.2",
        "libpq": "12.15",
    },
    python="3.9.13",
)
class NeuralNetFlow(FlowSpec):
   ..........

Any help or suggestions will be helpful.

hello. a few things: • I suppose you are using

@conda_base

since openssl and libpq are not pypi packages — the error line also seems to hint at that. •

METAFLOW_CONDA_DEPENDENCY_RESOLVER

has no effect unless you install the netflix metaflow extensions. • your environment seems overconstrained. I don’t know your exact requirements but this works:

dependencies:
  - tensorflow = 2.11.0
  - sentence-transformers = 3.0.1
  - matplotlib
  - pandas = 1.5.3
  - numpy = 1.25.2
  - psycopg2-binary = 2.9.9
  - transformers = 4.41.2
  - python = 3.9.13

(it installs matplotlib 3.3.2 so not sure if you absolutely need 3.4.2 — if you do, there may be other ones you can try floating to see).

Thanks @dry-beach-38304, Switched to

@pypi_base

and its working.