Delta sharing Features-Share live data directly - Easily share existing, live data in your Delta Lake without copying it to another system.Support diverse clients - Data recipients can directly connect to Delta Shares from Pandas, Apache Spark™, Rus...
For example would you use a feature store on your raw data or what's is the granularity of the features in the store?
I'll try to answer the broad question first, followed by the specific ones.When would you use the Feature Store?A Feature Store is primarily used to solve 2 challenges.(1) Discoverability and governance of featuresChallenge: In a large team or organi...
Yes.Please see Blog1: https://databricks.com/blog/2020/06/03/customer-lifetime-value-part-1-estimating-customer-lifetimes.htmlNotebook1:https://databricks.com/notebooks/CLV_Part_1_Customer_Lifetimes.htmlBlog2: https://databricks.com/blog/2020/06/17/c...
I am using ML flow and my need of the hour is to delete an experiment and want to create another experiment with same run.client = MlflowClient(tracking_uri=server) client.delete_experiment(1)This deletes the experiment, but when I run a new experim...
SQL Database:This is more tricky, as there are dependencies that need to be deleted. I am using MySQL, and these commands work for me:USE mlflow_db; # the name of your database DELETE FROM experiment_tags WHERE experiment_id=ANY( SELECT experime...
I'm a data scientist creating versioned ML models. For compliance reasons, I need to be able to replicate the training data for each model version. I've seen that you can version datasets by using delta, but the default retention period is around 30 ...
Delta, as you mentioned has a feature to do time travel and by default, delta tables retain the commit history for 30 days. Operations on history of the table are parallel but will become more expensive as the log size increasesNow, in this case - s...
Yes.Please refer to our docshttps://docs.databricks.com/applications/machine-learning/manage-model-lifecycle/multiple-workspaces.html
Yes!You will have to pip install mlflowin your environment as a first step. For more details, see: https://docs.databricks.com/applications/mlflow/access-hosted-tracking-server.html
How does it provide a glass box view?
Depending on which solution you use, GlassBox means that any interactive work you do via point & click, we automatically generate the code behind the scene and generate notebooks used for each experiment that was ran under the hood, in addition for a...
We have been using open source MLflow, how will it benefit us to move to Databricks mlflow?
Please see https://databricks.com/product/managed-mlflow
AutoML presumably tries a few different algorithms while hyperparameter searching. What model types are considered?
At the moment, it's really just xgboost, and sklearn implemenations like random forests, logistic regression, and linear regression as applicable. More possibilities are coming.
I have an NLP application that I build on my local machine using spacy and pandas, but now I would like to scale my application to a large production dataset and utilize the benefits of sparks distributed compute. How do I import and utilize a librar...
It depends on what you mean, but if you're just trying to (say) tokenize and process data with spacy in parallel, then that's trivial. Write a 'pandas UDF' function that expresses how you want to transform data using spacy, in terms of a pandas DataF...
I don't think we have a lot of internal docs, just high-level explanations like https://databricks.com/blog/2021/05/27/databricks-announces-the-first-feature-store-integrated-with-delta-lake-and-mlflow.htmlHowever I don't think there's much to it. Th...
The feature store has both online / offline components. The offline feature store is used for feature discovery, model training, and batch inference and is backed by Delta tables. You could read/write to offline store from Databricks clusters that...
Best NLP APIs to use with Spark which gives better performance
By far the most popular and comprehensive library, to my knowledge, for Spark-native distributed NLP, is spark-nlp from John Snow Labs. https://nlp.johnsnowlabs.com/ It is open source (but with commercial support options) and has a whole lot of funct...
