Today I will be participating in a discussion on the opportunities and pitfalls of various trending methods of connecting data to AI, along with Jeff Huber founder of Chroma , Jeffrey Morgan creator of Ollama, and Jorge Torres of MindsDB

OpenBB, MindsDB, Nixtla 🤝

🎉🚀📈 New release for NeuralForecast 📈🚀🎉 At Nixtla we are very excited to share the latest release of NeuralForecast, including new models, features, and tutorials 🤖💡 ⏰ Old meets New. Two new exciting models: DeepAR and TimesNet. DeepAR is one of the earliest neural

🚀Thrilled to announce nixtla participation in the "The New FinAI Tech Stack" event at MindsDB! Join me on 12/7 for an exclusive showcase of AI advancements in Finance. Save the date and let's dive into the future of financial tech! 🌐✨ lu.ma/84m98jgo

About time series, Clement Delangue said: "Recently, Time GPT has been launched, and we're witnessing a growing number of startups exploring this area. What's particularly exciting is the potential to integrate time series with other modalities." 1/7

🔥 TL;DR: Lag-Llama is 42% less accurate and 1000x slower than a simple seasonal naive model. 😲 🚀 We are extremely happy to see other companies following in nixtla's footsteps in creating foundational models for Time Series. Recently, we explored LagLlama, a model proposed

It was great to host a session at Time Series Analysis And Forecasting Society (TAFS) yesterday where Max Mergenthaler and azul (she/her) of nixtla talked about "TimeGPT: A Foundation Model for Time Series." In the evolving landscape of open-source time series Large Language

Have you wanted to learn more about TimeGPT, our pre-trained generative model for time-series data forecasting? On HackerNoon | Learn Any Technology, Jorge Torres of MindsDB puts TimeGPT under the microscope, analyzing it from all angles, to show how you can integrate it with your database to

🎉 TL:DR: Amazon's new foundational model Chronos is 10% less accurate and 500% slower than training classical statistical models 🤯 🤔 Six months after we published the first foundational model for time series, nixtla welcomes Amazon Science to the party. 🥳 🔥 We

🎉 TL:DR: Salesforce Moirai performs well in hourly data and is much faster than Amazon Science Chronos but is still up to 33% less accurate and efficient than statistical models 🤯 🤔 🔥 nixtla presents a comprehensive, reproducible evaluation demonstrating that a

New Release to NeuralForecast 🚀 by nixtla ✅ BiTCN - temporal convolutional networks forecasting model ✅ iTransformer - transformer-based forecasting model ✅ MLPMultivariate - an MLP model that supports multivariate tasks

How to pick the best-performing time-series AI model for your specific data? MindsDB answers just this question in their article. There are many AI frameworks for time-series forecasting, but they may perform differently on different types of data. How to determine which model

Improve large-scale, time-series forecasting with nixtla's latest Lightning AI Studio. Take advantage of the capabilities of NeuralForecast, an open-source Python library, and efficient compute, parallelization, and hyperparameter tuning to generate accurate predictions for

it is difficult not to get a little bit emotional after seeing Satya Nadella present TimeGEN-1 on Microsoft Azure. i feel very proud of the team i'm honored to be a part of at nixtla. 💙

TLDR: TimeGPT-1 by nixtla is the most accurate and fastest among foundation models (closely followed by Google AI's TimesFM). 🚀🚀🚀 We heard you. And we now present a reproducible benchmark comparing different foundation and classical models on 30K unique series. We

We’re excited to introduce @ChaiDiscovery and release Chai-1, a foundation model for molecular structure prediction that performs at the state-of-the-art across a variety of drug discovery tasks We're releasing inference code, weights & a web interface: chaidiscovery.com/blog/introduct…