This message was deleted BentoML #ask-for-help

Join Slack

This message was deleted.

# ask-for-help

Slackbot

05/02/2023, 5:09 PM

This message was deleted.

Fernando Camargo

05/02/2023, 9:28 PM

I managed to reproduce the problem locally by running the commands using Docker:

Copy code

docker run --rm -v $PWD:/workspace -w /workspace -e HOME=/workspace -e PIPENV_VENV_IN_PROJECT=true <http://gcr.io/klue-dev-310020/pipenv:python-3.10|gcr.io/klue-dev-310020/pipenv:python-3.10> pipenv run bentoml build -f src/deployment/torch/bentofile.yaml

I ran this command right after running the previous one the packs the model. Since I'm mounting the current dir as

/workspace

and setting it as the

$HOME

, I can see the

bentoml

folder, containing the model there.

Fernando Camargo

05/02/2023, 9:36 PM

I added some logging messages and discovered that it's getting stuck on this portion of the code:

Copy code

for dir_path, _, files in ctx_fs.walk():
            for f in files:
                path = fs.path.combine(dir_path, f.name).lstrip("/")
                if specs.includes(
                    path,
                    recurse_exclude_spec=filter(
                        lambda s: fs.path.isparent(s[0], dir_path),
                        specs.from_path(build_ctx),
                    ),
                ):
                    target_fs.makedirs(dir_path, recreate=True)
                    copy_file(ctx_fs, path, target_fs, path)

Fernando Camargo

05/02/2023, 9:48 PM

Solved the problem by changing my

bentofile.yaml

from:

Copy code

include:
  - "*.py"

to:

Copy code

include:
  - "src/**/*.py"

The problem was that it was looking for all the .py files possible (including the ones inside

/.venv

3 Views

Open in Slack

Previous Next