Meta Open Source的动态

"Week 5 of my #100WeeksOfAI challenge is a wrap! ?? This week, I tackled building an LSTM from scratch, diving into sequence modeling and understanding how neural networks handle memory for complex tasks like language processing. You can find the implementation on my GitHub: [https://lnkd.in/g2nCXEnh ]. ???? Excited for the upcoming challenges as I continue implementing a new AI research paper each week. Stay tuned for more! ?? #AI #MachineLearning #DeepLearning #LSTM #SequentialData #100WeeksOfAI"

?? Exploring the Depths of Perceptrons: From Single-Layer to Multi-Layer Architectures! ?? Dive into the fascinating world of neural networks with me as we unravel the mysteries of single-layer and multi-layer perceptrons! ?? ?? Check out my latest project on GitHub, where I delve into the implementation and understanding of Perceptrons: ?? GitHub Link: Perceptron Implementation From basic concepts to practical implementations, this journey promises to expand your horizons in the realm of artificial intelligence and machine learning. ?? Ready to embark on this adventure? Let's explore the boundless possibilities of Perceptron together! ?? #MachineLearning #NeuralNetworks #ArtificialIntelligence #GitHub #Perceptrons #AI_Cadmy

After diving deep into the world of Machine Learning, I've just completed and uploaded my first Recommendation System project on GitHub! ???? This journey has been a rollercoaster of learning, from understanding algorithms to fine-tuning models. I'm thrilled to see how far I've come and excited for what lies ahead! ???? Check out the project here: https://lnkd.in/d4fsDZWS ?? #MachineLearning #AI #DataScience #RecommendationSystem #LearningJourney #GitHub

Go watch this video to know about why you have to learn pythone in 2024 and perhaps this will be continued in trend further in future because with pythone you can do many things in any field like Data science Machine learning And ARTIFICIAL intelligence field I think learning pythone In 2024 is best choice for everyone who want to be in THis AI Revolution field to be more productive... share your thoughts in comment?? https://lnkd.in/dRb-hfd9

Just finished this course. This is a good foundation to my Data Science, AI/ML journey.

Learn how you can fine-tine every ML algorithm in Python by using open-source framework Optuna: Cristian Leo's guide provides all the guidance (and code) you'll need to get started.

Thank you DeepLearning.AI for creating such a simple and easy-to-follow course. I loved the intuitive interface, where you can run notebooks and videos side by side. It was a good refresher for me. #lifelonglearner #pythonprogramming #llm #ai #codingcompanion

I learn how to make a AI application by flask and how to deploy it. The application can have an error handling and already pass static code check.

?? AI Learning Day #3 ?? Today I've run a successful "hello world" on LangGraph, then I've reproduced a fully mocked workflow to better understand what's going on under the hood.?? I’m excited to share my progress and hope it’s useful for someone out there! ?? ?? Check out my detailed notebook on GitHub: https://lnkd.in/dq9SGkzF #AI #MachineLearning #LangChain #LangGraph #PythonDevelopment #JupyterNotebooks #CodeNewbie #DeveloperCommunity #DataScience #TechInnovation

#Day68 #GenerativeAI #75Hard Meta LLAMA 3 Fine Tuning using ORPO is Here ?? ?? Video Link - https://lnkd.in/d5rD3pD6 ?? In this video I Explained about - 1. What are the new improvements in Meta LLAMA 3 over LLAMA 2 2. What is ORPO Fine Tuning and How it Work and Why it is best? 3. How ORPO Fine Tuning Works with Maths and Formula. 4. What Kind of Data is used in ORPO Fine Tuning and How to Tokenize it? 5. Load Meta LLAMA 3 Model using Hugging Face Library. 6. Load the LLAMA 3 in 4 Bit Precision using QLORA Configuration and Bitsandbytes Library. 7. PreProcess the Dataset. 8. Train the Model and Save it to the Hugging Face Repository. 9. LLAMA3 Model Evaluation using LLM-AutoEval after Fine Tuning. ORPO is a new exciting fine-tuning technique that combines the traditional supervised fine-tuning and preference alignment stages into a single process. This reduces the computational resources and time required for training. ORPO modifies the standard language modeling objective, combining the negative log-likelihood loss with an odds ratio (OR) term. This OR loss weakly penalizes rejected responses while strongly rewarding preferred ones, allowing the model to simultaneously learn the target task and align with human preferences. Meta LLaMA 3: Two model sizes have been released - a 70 billion parameter model and a smaller 8 billion parameter model. The 70B model has already demonstrated impressive performance, scoring 82 on the MMLU benchmark and 81.7 on the HumanEval benchmark. ORPO requires a preference dataset, including a prompt, a chosen answer, and a rejected answer. In this example, we will use "mlabonne/orpo-dpo-mix-40k". I use Llama 3 8B model in 4-bit precision with LORA Configuration and PEFT Library to load the Model in small size. ?? Kaggle Notebook Link - https://lnkd.in/dTBcHUMx ? Get Started with GenAI for Free [ Day 1 to Day67] - https://lnkd.in/dpCpTXUz ?? Get 1:1 Mentorship or Career Guidance in GenAI, LangChain or Data Science and Machine Learning - https://lnkd.in/d3r6wAMz Simranjeet Singh ?? Follow to Learn AI ??