deepset-ai / FARM

:house_with_garden: Fast & easy transfer learning for NLP. Harvesting language models for the industry. Focus on Question Answering.
https://farm.deepset.ai
Apache License 2.0
1.73k stars 247 forks source link

Does the public MLFlow server support uploading artifacts and models? #831

Closed Ark-kun closed 2 years ago

Ark-kun commented 3 years ago

First of all, thank you a lot for providing public MLFlow server. MLFlow is surprisingly inconvenient to deploy, so this really helps trying out farm.

I'm trying to get a grip on model logging and it looks like I cannot log any artifacts.

from farm.utils import set_all_seeds, MLFlowLogger
ml_logger = MLFlowLogger(tracking_uri="https://public-mlflow.deepset.ai/")
ml_logger.init_experiment(experiment_name="Public_FARM", run_name="Run_doc_regression")
ml_logger.log_artifacts(mlflow_model_path, artifact_path="Artifact_1")
07/18/2021 02:22:25 - WARNING - farm.utils -   Failed to log artifacts: Failed to upload /home/jupyter/mlflow_model.MLFlowModel/conda.yaml to public-mlflow/2/da79489fac914ae7b2f59cf9691e57d2/artifacts/Artifact_1/conda.yaml: An error occurred (AccessDenied) when calling the PutObject operation: Access Denied

Is this expected?

Timoeller commented 3 years ago

Our public MLFlow instance is rather small, so logging artifacts or even whole models would made it crash.

What kind of problems arise when setting up your own MLFlow, maybe also the MLFlow community help there?

stale[bot] commented 2 years ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 21 days if no further activity occurs.