Kamal Dhungana

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

🎉 29 Days of Airflow 2.9: Day 28 (2 of 3) 🎉 🔧 Today in Apache Airflow 2.9, we're rolling out a crucial tool for workflow management: adding CLI support for bulk pause and resume of DAGs. This enhancement allows users to pause or resume multiple #DAGs at once via the command line. This is particularly useful for managing large numbers of DAGs efficiently, providing a significant boost in operational flexibility. ⚙️ Why does this matter? Bulk pause and resume capabilities mean that users can quickly make broad adjustments to their workflows, saving time and reducing the complexity of operations. This is especially valuable during system maintenance, updates, or when adapting to sudden changes in data processing needs. 🔗 Learn more about how bulk pause and resume commands can streamline your DAG management: https://bit.ly/455tNxh #Airflow29 #29daysofairflow29 #DataEngineering #WorkflowManagement #Airflow

3

🚀 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

When it comes to "Retrieval" phase of RAG, choosing the right strategy makes all the difference. Sparse retrieval (like BM25/TF-IDF) is great for finding exact keyword matches, while Dense retrieval (like BERT) understands the meaning behind words for more accurate results. Hybrid retrieval combines both, offering the best of both worlds. Each method has its strengths. Understanding when to use them is key to better search results. 

10

4 Comments

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

Overfitting in ML: Why It Matters Overfitting happens when a model is too focused on training data, causing poor real-world performance. Stay aware to get more accurate predictions. Read more : https://lnkd.in/gNcyi_x #datascience #machinelearning #overfitting #ai #bigdata #mlalgorithms #techinsights #deeplearning #artificialintelligence #datamodeling #dataanalytics #linkedin #youtube #twitteonx #threads #instagram #like #share #sprintzeal #sprintzealcommunity

2

1 Comment

📊 Optimizing RAG Applications with TruLens: A Quick Guide Recently explored TruLens for evaluating and fine-tuning Large Language Models (LLMs) in Retrieval Augmented Generation (RAG) applications. Here's what I learned: Key Evaluation Metrics: ✅ Groundedness: Ensures responses are factually supported by context ✅ Answer Relevance: Measures if responses directly address queries ✅ Context Relevance: Evaluates quality of retrieved information The Process: Set up automated evaluation pipelines using TruLens Track performance across different configurations Use metrics to optimize: Chunking strategies Retrieval parameters Context window sizes Embedding models Benefits: Data-driven optimization Automated quality assessment Systematic performance tracking Clear metrics for improvement Result: More accurate, relevant, and reliable RAG applications with measurable improvements in response quality. #AI #MachineLearning #LLM #RAG #DataScience #TruLens https://lnkd.in/gHQnjREi

3

✨ Exciting Side Project Discovery: LLM + MCTS Integration! ✨ Arushi Somani recently embarked on an intriguing project combining Large Language Models (LLMs) and Monte Carlo Tree Search (MCTS). Here's the setup: She utilized Humaneval and LLaMA. Implemented MCTS-assisted inference decoding with Humaneval pass/fail as the reward and performed k rollouts per question. Compared the results to sampling the model for pass@k scores. Her findings: For sufficiently large k, MCTS-assisted inference consistently outperforms random sampling. On average, while random sampling needed 10 tries to get the correct answer, MCTS only required 6.77 attempts. over enough attempts, MCTS searches solution space more efficiently Writeup : https://lnkd.in/dB82vYc4 Code : https://lnkd.in/ddUjmXzx References : https://lnkd.in/d3-QtmJ9 #AI #MachineLearning #LLM #MCTS #Innovation #DatamindAI #AIResearch

1

#KnowledgeGraph meets #LargeLanguageModels. Interesting #research on GNN-RAG by University of Minnesota. First, a GNN reasons over a dense KG subgraph to retrieve answer candidates for a given question. Second, the shortest paths in the KG that connect question entities and answer candidates are extracted to represent KG reasoning paths. The extracted paths are verbalized and given as input for LLM reasoning with RAG. #GenerativeAI #Innovation #RetrievalAugmentedGeneration #GNN Full paper: https://lnkd.in/dkqhVHnC

16

Staying updated with ML research is challenging. Our latest blog demonstrates how to build a bot summarizing arXiv papers using W&B Weave and Anthropic's API. Check out the technical details and follow along with our interactive Colab. Read more: https://lnkd.in/ggratqF7

24

The Science of Winning — A reminder of the history of O.R. and how far it’s come [OR/MS Today] https://bit.ly/44Oelpv #informs #orms #operationsresearch #optimization #simulation #AI #ML #datascience #analytics #managementscience

4

Webinar alert - April 30, 1PM EDT. George Anadiotis from LinkedDataOrchestration and George Demarest of #Aerospike will guide you through selecting the right graph database and essential selection criteria tailored to your organizational needs. https://lnkd.in/eDV-QQuB #GraphDatabase #DataManagement

10

GeoPandas offers an intuitive way to create interactive maps through its explore() method. In this tutorial, we’ll dive into how to use this method to…  #statology

1

A clean (devoid of unnecessary abstractions) experimental package released by OpenAI for multiagent systems. There is no official support as of now to the package, but few interesting examples. Also like their simple take on defining agents - An Agent simply encapsulates a set of instructions with a set of functions (plus some additional settings below), and has the capability to hand off execution to another Agent. MIT license. Good read for exploration. Link in comments. [openAI, agents, LLMs]

11

2 Comments

🧠 Simplifying the Complex: SelMix Optimizing complex performance metrics in machine learning has long been a challenge for researchers and practitioners alike. Traditional methods often fall short when dealing with real-world datasets and objectives. That's where SelMix comes in. Our research, in collaboration with the Indian Institute of Science (IISc), introduces SelMix (Selective Mixup), a fine-tuning method that optimize complex performance measures for classification efficiently. SelMix fine-tunes pre-trained models through selective mixup, significantly improving the optimization of non-decomposable objectives in both semi-supervised and supervised learning settings. This method excels at handling imbalanced datasets (long-tailed datasets) and ensures models improve classification performance across all classes, especially for minority ones.🔧 In fact, SelMix demonstrates superior capability in both semi-supervised and supervised learning setups, significantly outperforming state-of-the-art approaches like CSST, DASO, and MiSLAS. 📘Join us in celebrating this achievement and delve deeper into our spotlighted paper at ICLR 2024 by reading the full #TechBlog: https://lnkd.in/gjnZPKfG (日本語), https://lnkd.in/ghunsieE (Eng) #ICLR2024 #MachineLearning #IISc #SelMix

27