Duncan McKinnon

🚀 Want to employ the power of #PyTorch for your scientific models without a steep learning curve? 👀 Check out our latest blog, where Travis Oliphant discusses how PyTorch's integration with the Array API makes it easy to use with familiar libraries like NumPy and scikit-learn.🔥 https://lnkd.in/dDRkEY3v #Innovation

10

Learn how you can convert a PyTorch model, trained on the web, to a mobile-optimized format. Madhur Zanwar describes the process using a few lines of code. He also discusses his journey and how one approach didn't yield satisfactory results while trying to run it in a React Native app. #machinelearning #pytorch #computervision #edgeml #reactnative #objectdetection #yolov5

14

There is often an unspoken assumption by ML researchers that all we need to make progress and model every aspect of our world is data. We do not ascribe to this in my lab. In fact, we believe that data will never be enough. Our experiences with the nuances and complexities of scientific problems have informed us to the insufficiency of data to capture continuous physical processes, which afterall govern our biological and physical world. An example of this is this series of two papers led by my wonderful PhD student, Anowarul Kabir and advanced by a precious multi-year collaboration of my lab with Los Alamos National Lab: Anowarul Kabir, Manish Bhattarai, Kim Rasmussen, Amarda Shehu, Anny Usheva, Alan R Bishop, and Boian S Alexandrov. Examining DNA Breathing with pyDNA-EPBD. Bioinformatics 39(11):btad699, 2023. https://lnkd.in/gZaHE4hS Anowarul Kabir, Manish Bhattarai, Selma Peterson, Yonatan Najman-Licht, Kim Ø Rasmussen, Amarda Shehu, Alan R Bishop, Boian Alexandrov, Anny Usheva. DNA Breathing Integration with Deep learning Foundational Model Advances Genome-wide Binding Prediction of Human Transcription Factors. Nucleic Acids Research: gkae783, 2024. https://lnkd.in/gfTTqcPy Our goal: advance an exceptionally challenging problem in molecular biology, prediction of transcription factor binding sites. Our first step: capture the underlying physics that is missing in the data. Our second step: integrate that now with the data we have in a foundation model for predicting transcription factor binding sites. Performance improves. Most importantly, when we look at the sequence motifs that constitute "signatures" of what makes a transcription factor binding site, we obtain answers. All in the open, nothing opaque.

33

1 Comment

Combining the power of TensorFlow, NeuralForecast, and zero-shot LLMs, Jinhang Jiang shows how we can generate accurate multi-step quantile forecasts.

11

Can you use #llms for every machine learning task? Members of MIT Data-to-AI Lab Kalyan Veeramachaneni and Sarah Alnegheimish tried #llms for a well known task of unsupervised time series anomaly detection in their paper. In an article in VentureBeat they explore the affordances that #llms provide and how to make most of them. Article: https://lnkd.in/emvEDDkg Research paper: https://lnkd.in/ezNzHkBj

5

How to fine-tune #LLMs on custom #datasets at Scale using Qwak and #CometML Alexandru Razvant https://lnkd.in/gP4aWM69

3

Working with large language models has become a critical part of any data scientist’s or ML engineer’s job, and fine-tuning the large language models can lead to powerful improvements in the language models’ capabilities. Eivind Kjosbakken's article shows you how you can fine-tune Llama3.2 to improve its performance within a targeted domain. #LLM #Llama

13

Rohit Patel's article explains how Large Language Models (LLMs) work, from scratch — assuming only that you know how to add and multiply two numbers. Check out the full guide now! #LLM #DataScience

40

1 Comment

Fine-tuning LLMs with Reinforcement Learning from Human Feedback This blog post explores how Reinforcement Learning from Human Feedback (RLHF) can be used to fine-tune Large Language Models (LLMs) to align with human preferences. It delves into the concept of Proximal Policy Optimization (PPO) and how it is used in this context. The key takeaway is that RLHF offers a powerful framework to train LLMs that can produce human-aligned text, while still maintaining creativity. This paves the way for more natural and nuanced language generation in the future of AI and natural language understanding. Read the article: https://lnkd.in/d9EGf579 #AI #machinelearning #goML

8

This is a really insightful piece by Mauricio Anderson from X3M about orchestrating the training of online ML models with Apache Airflow! 🚀 Check out the article to learn how X3M leverages Airflow for efficient and scalable ML model training below! #ApacheAirflow #MachineLearning #DataScience #X3M #AI #TechInnovation

89

1 Comment

When *open* isn't *open source* 🙋🏼‍♂️👀 Reflecting on the open-source landscape for Large Language Models, I explore the essentials: permissive licensing, transparency in training data and model architecture & more #LLM #LLMs #AI #GenerativeAI #opensource https://lnkd.in/gSJuK_J4

1

Welcome to our new blog Series on Friendli Model Optimizer! Friendli Model Optimizer is a quantization tool that you can use to improve inference speed, lower resource consumption and maintain model accuracy. In this article, we’ll explore how to use the FMO library and understand its quantization feature. Stay tuned for part 2 of this blog, where we'll dive deeper into the performance analysis. Read Full blog 👉 https://lnkd.in/d-WU5MgJ Try friendli Suite 👉 https://suite.friendli.ai/ #FriendliAI #Friendli #FriendliModelOptimizer #LLM #Inference

14

New Blogpost: Fine-tuning LLMs for Multi-turn Conversations In this deep dive we cover: 🎛️ How to fine-tune with conversation data. 🎭 Implement loss masking to improve model performance. 🐍 Full code example fine-tuning Llama 3.1 8B on CoQA dataset. Read here: https://lnkd.in/gFxbpia6

20

Explore how to fine-tune BERT using the Hugging Face Transformers library. Discover how you can leverage pre-trained models for your own projects from Shaw Talebi's article. #LLM #MachineLearning

13

Only ✨15 DAYS✨ until PyTorch Conference! Join us in San Francisco, Sept 18-19, to help shape the future of PyTorch, engage in technical sessions, explore collaborative opportunities, and connect with a global network of developers. Check out the schedule: https://hubs.la/Q02NwT3D0 Grab your ticket today: https://hubs.la/Q02NwL-m0 #PyTorchConf #PyTorch #PyTorchFoundation #AI #MachineLearning #ML #DeepLearning #OpenSource #OpenSourceSoftware #OpenSourceDevelopment #OpenSourceCommunity #OSS #LinuxFoundation #events #LLM #NLP

12

Math often begins as a fun exploration, but its real-world applications are countless. In our latest blog post, we uncover lesser-known math concepts that are closely related to data science, yet rarely featured in traditional curriculums. It's a useful post where we share accessible articles as well as in-depth resources. In other words: want to learn about math in data science? We've got you covered! Let's dive in 📈 #DataScience #MathInDataScience #ChaosTheory #FourierAnalysis

11

What's the difference between model explainability, interpretability, and observability? Jason Zhong's debut TDS article outlines the nuanced distinctions you should be aware of.

15

1 Comment

🔔 We are excited to introduce our first series of open-source #Small Language Foundation Models, Prem 1B and Prem 1B chat. These models are available on Hugging Face under APACHE LICENSE 2.0. 📍 Our goal is to create models that excel at #RAG. Since RAG works by processing information at runtime, the main constraint is LLM size. For RAG, models don't need to be huge; they just need strong text comprehension to give accurate answers when provided with the right context. 🥊 Prem 1B was pre-trained on SlimPajama dataset. We used Lightning AI and Ray by Anyscale to orchestrate our data processing and training pipeline. Training was done on 16 H100 GPUs distributed across 2 nodes. We trained our model for 2 epochs on 1.2 T tokens. 🌐 Prem 1B is a transformer-based SLM, similar to TinyLlama, and uses Flash Attention. We explored architecture styles such as State Space models like Mamba and MoEs, but learned that data quality eventually determines everything. Prem 1B supports context length up to 8192 tokens. 💭 For Chat-Finetuning (SFT), we trained the models with a dataset supporting multi-turn conversation. We used chat template and tokenizer from #Llama3. For our alignment step, we used DPO and popular datasets like UltraFeedback, Orca DPO Pairs, OASST2 DPO Pairs. Read our release blog here: https://lnkd.in/d4ZFB2Xi

38