Eugene K.

🚀 Looking to streamline fine-tuning for models like Llama 3, Mistral, and Mixtral? Our latest blog shows how Anyscale simplifies LLMs fine-tuning with Ray: 🔑 Key takeaways: • Scalable fine-tuning with LLM-Forge • Model Deployment with RayLLM Read more: https://lnkd.in/gQwD4A2U

34

🔍 Transforming Intent Classification: From TF-IDF to SBERT 🚀 In the world of NLP, intent classification is a key player in understanding and predicting user intents. As models evolve, moving from traditional TF-IDF vectors to SBERT-based embeddings has opened new doors to more accurate and context-aware predictions. Let’s dive into how TF-IDF and SBERT differ, and why SBERT is making a difference for NLP applications in intent classification! 👉 #NLP #IntentClassification #TFIDF #SBERT #MachineLearning #AI

4

Lin Quiao wanted to merge PyTorch with Caffe2 in a 'zipper approach', but the APIs didn’t align. Instead, their team switched to focus entirely on #PyTorch and helped make it a high-performance, production-ready framework. Discover more at #PytorchConf 2024, running September 18-19 in San Francisco. Register now: https://hubs.la/Q02NqL230 #Pytorch #opensource #PyTorchConf

14

Building Optimized RAG with LlamaIndex + DSPy 📈 We’re excited to announce a comprehensive set of integrations with DSPy that let you combine DSPy’s PyTorch-esque syntax and optimization capabilities with the comprehensive set of data+orchestration tools around RAG/agents that LlamaIndex offers. If you prefer to use DSPy Signatures to define your prompt+LLM input/outputs, you can do that. If you prefer to use LlamaIndex prompt modules and leverage DSPy optimization capabilities, you can do that. If you want to repurpose existing optimized DSPy predictors as LlamaIndex prompts, you can do that too 🔥 The three integrations are as follows: 1️⃣ Build and optimize Query Pipelines with DSPy predictors  2️⃣ Build and optimize Query Pipelines with LlamaIndex Prompts, but use DSPy to optimize 3️⃣ Port over DSPy-Optimized Prompts to any LlamaIndex Module with the `DSPyPromptTemplate` syntax. A bonus ✨: Learn how to define custom evaluator functions using LlamaIndex evaluators that still plug into the DSPy optimizer Thanks to both Omar K. and Arnav Singhvi for the help with this. Full cookbook here: https://lnkd.in/gCZjqui4 

412

13 Comments

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

11

🚀 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

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

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

Have you noticed how large language models (LLMs) are all the rage? For instance, Llama 3 by #Meta with 400 billion parameters! (Parameters in large language models (LLMs) are the learned #variables that define how the model interprets and generates language.) But guess what? Tech giants are shifting their #focus to Small Language Models (SLMs).  Curious about what they are? Let’s dive in!  #ai #llm #slm #machinelearning #ml #innovation #techtrends #GenAI #trends #opportunity #new #microsoft #apple #appleintelligence #microsoftphi

17

"Learn Your Reference Model for Real Good Alignment," introduces Trust Region Direct Preference Optimization (TR-DPO). Developed by a visionary team led by Alexey Gorbatovski at Tinkoff, this paper pioneers a method that significantly advances the training of language models by updating the reference model during the training process itself. Why This Matters: ➡️ Enhanced Model Alignment: TR-DPO introduces dynamic updates to the reference model, which allow for more accurate and adaptive training compared to the traditional Direct Preference Optimization (DPO). ➡️ Improved Performance Metrics: Through rigorous testing, TR-DPO has demonstrated a 19% improvement in performance over DPO, making it a groundbreaking tool for achieving higher coherence, correctness, and detail in model outputs. Key Insights: ➡️ Dynamic Reference Model Updates: Unlike static methods, TR-DPO continuously adjusts the reference model during training, which keeps the model relevant and highly effective across changing data sets and conditions. ➡️ Human-Centric Results: The model notably excels in producing outputs that are coherent, detailed, helpful, and harmless—key qualities for any application from chatbots to complex decision-making systems. #analyticsvidhya #datascience #generativeai #researchpaper

16

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

Ever needed fast inference and cost-effective fine tuning? Try out FriendliAI’s inference endpoints or fine-tuning tools. As of release 1.26.3, Weaviate has an integrated Generative Search Module to FriendliAI hosted models. Find out more: https://lnkd.in/gQtqgdHx

20

Build powerful RAG pipelines in Friendli Suite with Weaviate database to generate personalized and context-aware responses!

15

Build robust GenAI pipelines with LlamaIndex, Qdrant, and MLflow for advanced RAG! Learn how to: 🔍 Streamline RAG workflows for better scalability 📊 Ensure performance consistency across model versions 🚀 Optimize indexing systems for efficiency This step-by-step guide covers: • Setting up MLflow for experiment tracking • Configuring Qdrant for vector storage • Integrating Ollama for LLM and embedding models • Creating a playground app to test the full pipeline Check out the full post here: https://lnkd.in/gCnm9JgE

230

5 Comments

Hyperparameter optimization and other tuning workflows (for Python or other programs) scale well when done in Chapel. Check out this blog post for how to do it: https://lnkd.in/gc_CZiG5 scikit-learn Kaggle fast.ai

5

The NeurIPS Navigator: A Behind the Scenes Look, Powered by Cerebras Inference ⚡ How do you create RAG app that turns 4,500+ research papers from the 2024 NeurIPS Conference? Dive into how we processed PDFs, created embedding chunks and stored it in on Postgres in a performant way. Created by our Fellow Isaac Tai Read more here: https://bit.ly/4fUOf8m

59

2 Comments

Learn to build an automated research paper report generator using LlamaIndex that downloads papers from arXiv, processes them with LlamaParse, and indexes them in LlamaCloud! Report generation is an increasingly important use-case so we're doubling down on making it easy. In this demo from Ravi Theja Desetty you'll learn how to: ➡️ Generate queries based on a report outline and classify them for LLM vs index-based answers ➡️ Handle paper downloading, metadata extraction, and document parsing automatically ➡️ Use Pydantic models to extract key metadata like author names, companies, and AI topics ➡️ Create structured reports with sections & subsections filled from both LLM and indexed content It supports customizable report outlines and templates for different use cases. Check it out here: https://lnkd.in/gCXKzeuH

140

2 Comments