2023 December Newsletter

January 17, 2024 · less than a minute read
Frame 70326

Make sure to sign-up for our monthly newsletter to receive all the latest AI and LLM news, tutorials and product updates.

Happy holidays from Predibase! It has been an undeniably exciting year for AI, and we’re happy to share the first edition of our newsletter, Fine-Tuned. In this edition we look back at some of our best-attended webinars and most-read blog posts as well as share a few recent exciting product updates including support for fine-tuning and serving Mixtral-8x7B.

Going forward, this newsletter will explore emerging best practices for building production AI, share hands-on tutorials, invite you to upcoming webinars and events, and highlight updates to the Predibase platform and our open source projects Ludwig and LoRAX.

Happy New Year!


Featured Event

Fine-Tuning Zephyr-7B to Analyze Customer Support Call Logs

Join us on February 1st at 10:00 am PT to learn how you can leverage open source LLMs to automate one of the most time consuming tasks of customer support: classifying customer issues. You will learn how to efficiently and cost-effectively fine-tune an open source LLM with just a few lines of code at a fraction of the cost of using a commercial LLM and how to easily implement efficient fine-tuning techniques like LoRA and quantization.

Featured Event

Recent Events + Podcasts

LLMOps for Your Data: Best Practices to Ensure Safety, Quality, and Cost
WEBINAR

LLMOps for Your Data: Best Practices to Ensure Safety, Quality, and Cost

Predibase CTO Travis Addair joined Shreya Rajpal, Co-Founder and CEO at Guardrails, to discuss how to work with LLMs in ways that mitigate risks, improve the development process, and better leverage your data as an asset.

Read full story
PODCAST

TFiR “Let’s Talk”: Predibase Makes It Easy For Anyone To Train LLMs

Predibase CEO Devvret Rishi joins the TFiR “Let’s Talk” show to share how developers can fine-tune open-source models with their own data to achieve significant cost savings and performance improvements compared to commercial LLMs.

Watch
PODCAST

Caveminds Podcast: Deploy Faster, Cheaper, Smaller AI Fine-Tuned Models

The discussion explores Predibase’s unique approach to AI, focusing on its ability to streamline complex AI pipelines into concise, efficient code.

Watch
Crafted | The Artium Podcast: Taking GenAI From Prototype to Production
PODCAST

Crafted | The Artium Podcast: Taking GenAI From Prototype to Production

This episode of “Crafted” is a recording of a panel Predibase Staff Engineer Justin Zhao spoke on at NYC Tech Week 2023 and covers how to take generative AI from prototype to production, how to overcome the challenges of evaluation and performance, and more.

Read full story

Featured Blog Post

AI and LLM Predictions for 2024

From the coming wave of small language models to the future of fine-tuning and LLM architectures, these predictions represent the collective thoughts of our team of AI experts with experience building ML and LLM applications at Uber, AWS, Google, and more.

Featured Blog Post

From the Predibase Blog

Graduate from OpenAI to Open-Source: 12 best practices for distilling smaller language models from GPT
Predibase Blog

Graduate from OpenAI to Open-Source: 12 best practices for distilling smaller language models from GPT

As a follow-up to our ,[object Object],, we’ve released an in-depth guide covering 12 essential best practices for distilling smaller language models from GPT. We’ve heard from plenty of customers and practitioners that commercial LLMs like GPT-4, while great for prototyping and proofs of concept, suffer from high costs and latency that often make them unsuitable for production applications. Read this post to learn how smaller, fine-tuned open-source models can help overcome these challenges.

Read full story
Fine-Tuning Zephyr-7B to Analyze Customer Support Call Logs
Predibase Blog

Fine-Tuning Zephyr-7B to Analyze Customer Support Call Logs

A typical customer support call costs an organization between $7 and $41, an expense that can quickly add up at scale. This tutorial–complete with an accompanying notebook you can follow along with–will teach you how to fine-tune an open-source LLM to accurately classifying customer support requests with just a few lines of code.

Read full story

From the Community

Large Language Model Fine-tining - Qlik Dork
Community Blog

Large Language Model Fine-tining - Qlik Dork

Follow along to learn how one user leveraged his Predibase free trial experience to fine-tune Llama-2-13b to accurately generate results for a made-up coding language.

Read full story
How to Fine-Tune LLMs without coding?
Community Blog

How to Fine-Tune LLMs without coding?

Learn how a member of the Ludwig open-source community used Ludwig to fine-tune Llama-2-7b.

Read full story
TechTalks: How to run multiple fine-tuned LLMs for the price of one
Community Blog

TechTalks: How to run multiple fine-tuned LLMs for the price of one

Read TechTalks’s coverage of LoRAX, the open-source framework Predibase released to enable teams to serve 100s of fine-tuned LLMs from a single GPU.

Read full story

Open Source Updates

Tutorial: How to Fine-tune Mixtral 8x7b with Open-source Ludwig

Mixtral 8x7B is one of the first successful open-source implementations of the Mixture of Experts architecture (MoE) and we’ve made it easy to fine-tune for free on commodity hardware using Ludwig—a powerful open-source framework for highly optimized model training through a declarative, YAML-based interface. Follow along with this hands-on tutorial!

Open Source Updates

Featured Product Update

We’re excited launch our new prompting experience in the Predibase UI which allows you to easily prompt serverless endpoints and your fine-tuned adapters without needing to deploy them first. This lets teams test their fine-tuned models and compare model iterations all from the UI, enabling much faster test and review cycles.

Featured Product Update

Full Product Updates

Inference Endpoints:

Predibase now offers instant access to Serverless LLM’s billed on a $/1k-tokens model as part of its Inference Endpoints. To see a full list of the serverless deployments available, visit our docs or our pricing page. Note: We’re constantly adding support for more models, please reach out to support@predibase.com with any requests.

Fine-tuning and Serving OSS Models:

With Predibase, you can now fine-tune and deploy any OSS LLM from HuggingFace up to 70B parameters with ease. Train state-of-the-art models via our fully-featured Python SDK or our intuitive UI and enjoy complete observability into your deployments afterwards.

Dedicated Compute:

Predibase now offers dedicated A100 capacity available on-demand. If you’re looking for access to state-of-the-art GPU’s for training or serving, contact us.

LoRAX New Release:

Predibase released LoRA Exchange (LoRAX) just a few months ago. Since then, we’ve added support for new models including Llama, Mistral, GPT2, Qwen, Mixtral, and Phi as well as new quantization techniques including bitsandbytes, GPT-Q, and AWQ. Stay tuned for even more exciting updates!

Want to try fine-tuning and serving LLMs on the most efficient, cost effective and easy-to-use AI platform out there? Then try Predibase for free with our trial!

Related Articles

Join Our
Slack Community!