Uptrain

Fixing an important production bug.

🚀 New UpTrain Release 🚀

This release includes improvements such as a better naming convention, regression accuracy measures, SHAP explainability and bug fixes.

* 📊 Added SHAP explainability as a feature: Users can now understand how their models work and make more informed decisions by using SHAP (SHapley Additive exPlanations) explainability.
* 📈 Added accuracy measures for regression tasks: Users can easily evaluate model performance on regression tasks.
* 📖 Added Readme for text summarization and fine-tuning LLM examples: Comprehensive documentation is now available for text summarization and fine-tuning LLM examples.
* 💻 Code cleanup and refactoring: Improvements such as fixing inheritance, better abstraction, and adding hover_text for graphs make it easier for users to work with UpTrain.
* 📝 Better naming convention, such as "monitors" instead of "anomaly": The improved naming convention provides a more accurate description of the feature's functionality.
* 🐞 Bug fixes: Several bugs were fixed in this release, improving the overall stability and performance of UpTrain.
* 📑 Added data integrity to text summarization example: The text summarization example now provides accurate and reliable results for users.

UpTrain has recently released an exciting new set of features to enhance its monitoring capabilities. One such feature is the addition of t-SNE dimensionality reduction and visualization, allowing users to gain deeper insights into their machine learning models. With t-SNE, users can now visualize high-dimensional embeddings in 2D or 3D, making it easier to spot patterns and anomalies in their data. This release also fixes a bug for using custom monitors with UpTrain statistics.

In addition, UpTrain now includes a finetuning for LLMs example, which demonstrates how to fine-tune large language models such as BERT and GPT-2 using UpTrain's monitoring framework. This example provides a step-by-step guide on how to set up and run a finetuning experiment. With this new example, UpTrain users can now easily adapt and improve their language models with confidence.

This release fixes some critical production-related bugs.

UpTrain announces a new feature release with the following features:

- Deprecated Tensorboard: UpTrain has replaced Tensorboard with Streamlit for a better and more user-friendly experience.
- Slack Alerts: UpTrain now offers Slack alerts for model monitoring, making tracking progress and staying informed about any issues easier.
- Customizable Port: With this update, users can now set the port number for the UpTrain dashboard according to their preferences.
- A/B Testing: UpTrain now offers functionality for monitoring A/B testing. This feature is particularly useful for businesses looking to compare the performance of different versions of their ML models.

UpTrain announces a new feature release with the following features:

* An example for text summarization task using Hugging Face.
* A new ability to visualize embeddings via UMAP.
* Normalization of vectors while clustering in data drift.
* Improvement of column names of measurables.
* Improved documentation.

These new features aim to improve the ability of the tool to help users with a wide range of machine learning tasks including text summarization in NLP. With the ability to visualize embeddings via UMAP, users can gain better insight into the structure of their data, while clustering can help to provide more accurate insights into data drift. The improved documentation and column names of measurables will make it easier for users to get started with the tool and understand its various features.

Overall, this feature release is an exciting step forward for UpTrain, and we look forward to seeing how it will help our users with their machine-learning observability.