External reviews
External reviews are not included in the AWS star rating for the product.
Great experience. It helped us create/deploy custom models on sagemaker.
What do you like best about the product?
The features. There are hell lot of features present on MLFlow
What do you dislike about the product?
Haven't explored everything. But yeah, maybe better documentation.
What problems is the product solving and how is that benefiting you?
Solving platform issues of aws sagemaker. We deployed a custom model which used to power ranking algorithm on the listing page.
- Leave a Comment |
- Mark review as helpful
Best Open-Source Platform to calibrate the models with keeping tracking of the experiments and store
What do you like best about the product?
One please with all nessarery fetaure.
1. tracking
2. storage of models and files related documentation like log, config file.
3. validation of the model with meteric feature and plot crossponding to it and genrate the report from the experiments.
4. Data callibation with data 🧱, SQL and other cloud providers.
1. tracking
2. storage of models and files related documentation like log, config file.
3. validation of the model with meteric feature and plot crossponding to it and genrate the report from the experiments.
4. Data callibation with data 🧱, SQL and other cloud providers.
What do you dislike about the product?
Anything which dislike is nothing tell yet, But If we can build something like feature where we can do more advanced anlaytics crossponding to parameters and meterics and generate the different plot from that tabuler data, like we have d-tale (github: https://github.com/man-group/dtale), beacuse I was running some simulation run with differnet experiments and then write the report crossponding with the experiments and with differnet signnificant plot, demonstrates the report write up.
What problems is the product solving and how is that benefiting you?
I was running some simulation run with differnet experiments and then write the report crossponding with the experiments and with differnet signnificant plot, demonstrates the report write up.
Good for code collaboration
What do you like best about the product?
Multiple people can write code in the same file at the same time. We use Databricks in Machine Learning.
What do you dislike about the product?
The cluster gets shut down after sometimes, leading to loss of data on the RAM
What problems is the product solving and how is that benefiting you?
Code collaboration, Machine Learning
A great tool for tuning your hyperparameters
What do you like best about the product?
One of the best utilities for packing code into reproducible runs and tuning the hyperparameters.
What do you dislike about the product?
Merge with Azure data bricks was not very smooth & giving problems while using GPU cores.
What problems is the product solving and how is that benefiting you?
I was making a recommendation system on Azure Databricks, & I tuned the hyperparameters of my model using the MLflow utility. It served the overall purpose and was giving satisfying results in A/B testing.
It was fantastic for all the integrations it had. Custom ml transform support could have been better
What do you like best about the product?
The fact that you can store all your models at one place
What do you dislike about the product?
The support for custom transforms isn't there
What problems is the product solving and how is that benefiting you?
We wanted a place to store all models at one place with versioning. Many benefits like serving the model are offered out of the box through databricks
Recommendations to others considering the product:
If you want to store all your ml models in one place then it's the way to go.
Perfect code sharing repository
What do you like best about the product?
Having a platform to share codebase with team members and run machine learning models on the cloud.
What do you dislike about the product?
Sometimes we have to restart clusters to fix memory errors, which leads to data loss.
What problems is the product solving and how is that benefiting you?
Collaboration on code among team members. Running applications on the cloud.
Brilliant on developing the best collaborative platform for data scientists and data engineers
What do you like best about the product?
An interface that is better than Jupyter notebooks that allows SQL, Scala, PySpark, Python, R and the ability to collabortate on notebooks
What do you dislike about the product?
DPU based billing is fixed and minimum is 3 node cluster. For a small entity the advantages of using AWS Glue interface to Spark 2.x outweighs the benefits of a permenant cluster runnig with Databricks.
What problems is the product solving and how is that benefiting you?
Big data management in lake type architecture using Parquet formats, PySpark developments and enhancements.
Databricks is the best option for your data workloads and pipelines
What do you like best about the product?
It is a highly adaptable solution for data engineering, data science, and AI
What do you dislike about the product?
I wouldn't say I like the lack of an easier way to import personalized code files or libraries from notebooks.
What problems is the product solving and how is that benefiting you?
I've solved emergency telephone data processing and insights. The performance of the solution is desirable.
Senior Cloud Evangelist and Architect
What do you like best about the product?
Spark Distribution of query and speed of batch query so does performance
What do you dislike about the product?
Interface can be make better and more intutive
What problems is the product solving and how is that benefiting you?
Big Batch bulk Parallel programming
Great platform for our Big Data needs
What do you like best about the product?
Easy administration, easy to create jobs from notebooks, great development environment, new and exciting features coming.
What do you dislike about the product?
Taking away our dedicated customer service rep and replacing this with just a support GUI.
What problems is the product solving and how is that benefiting you?
All our data pipelines are on Databricks. Benefitted from improved performance on Spark.
showing 321 - 330