Ramitha Manjunath

Many of ML practitioners might have heard the term "Extending context length" for LLMs. In case if you want to dive deeper into this topic along with its code implementation, give this blog a read. I have described about Position Interpolation, NTK aware and Dynamic NTK method, which are widely used for extending context length.

69

Check out this highly recommended course for those interested in optimizing their models: - Introduction to Quantization - Symmetric and Asymmetric Quantization Modes - Hands-on Implementation from Scratch using PyTorch - covering Per Tensor, Per Channel (Axis), and Per Group techniques - Weights Packing: Exploring 2 or 4 Bits Quantization Methods https://lnkd.in/gMAuWXHm

43

If you get a moment, please take this survey for this study on fairness in machine learning for my friend and colleague!

3

1 Comment

Excited to share that KerasHub is here! Whether you’re working in JAX, TensorFlow, or PyTorch, KerasHub ensures compatibility across backends, making it easy to plug into your existing workflows. If you’re looking to simplify your foundation model workflows, KerasHub might just be the tool you’ve been waiting for. #Keras #Google #NLP #LLMs

38

🌟 **Decoding Vector vs. Graph Databases for RAG & RAFT – Which One Wins?** When it comes to enhancing large language models through Retrieval Augmented Generation (RAG) or Retrieval Augmented Fine-tuning (RAFT), the choice of database—vector or graph—plays a pivotal role. 🔹**Vector Databases:** Data is partitioned into chunks and each chunk is transformed into a vector using a language model. These vectors are then indexed in a vector database, facilitating quick nearest neighbor searches to find the most relevant information chunks in response to a query. 🔸**Graph Database:** This database shines in storing relationships between entities. By parsing and indexing these connections, graph databases excel at constructing detailed subgraphs for specific queries, ensuring that only the most relevant, contextually connected information is retrieved. It’s about understanding the network of relationships.An LLM is good at extracting the below kind of triplet information. [Entity A] -> [Relationship/Context] -> [Entity B] Both have their strengths—vector databases are great for fuzzy matching large data sets quickly, while graph databases excel in precise context and relationship extraction. 🚀 **The cutting edge? Combining both!** : This hybrid model takes advantage of the quick searching capability of vector databases with the deep relational insights of graph databases. For RAG or RAFT, this means not only retrieving the most relevant information quickly but also understanding its context within a broader knowledge base. For example, if a model is answering a query about "healthy eating," a vector database can quickly fetch documents mentioning health and nutrition. A graph database then elaborates on how these concepts connect with diet plans, medical advice, and recipes, providing a comprehensive, context-rich response. #GraphRAG #RAG #RAFT #GenerativeAI #GraphDB #VectorDB

13

I am excited to share my latest Article on InsightGenix, an AI-powered SQL assistant that simplifies data analysis. Transform natural language into SQL effortlessly and gain valuable insights. Blog here: https://lnkd.in/g6xQU6tM Demo video: https://lnkd.in/gxrh3kXv Take your data analysis to the next level with InsightGenix! #InsightGenix #SQLAssistant #Streamlit #DataAnalysis #AI #OpenSource #DataInsights #AzureOpenAI #BlogUpdate

13

This short course is a great introduction to RLHF. Highlights: 1. Instructor explained what is RLHF, how it is used in LLMs to align responses to human preferences. 2. How to prepare preference datasets and do reward modelling . 3. How to use reward model in RL Loop to do fine tuning of LLM . All of the above was demonstrated on GCP vertex platform using pipelines and llama2-7B model on summarization task . Next logical learning adventure would be to go through trl library from hugging face , going through blogs and their smol course https://lnkd.in/gBKpJ7vS #rl #rlhf #llm #finetuning #learninginpublic

25

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

Function calling in Large Language Models (LLMs) is redefining how AI integrates with real-world systems. By enabling models to invoke external functions or APIs, this feature goes beyond generating text to seamlessly performing actions like fetching live data, triggering workflows, or executing specific tasks. For instance, imagine asking an AI to calculate complex formulas, retrieve weather updates, or automate scheduling—it’s all possible with function calling. This capability allows LLMs to generate structured outputs (e.g., JSON) that external systems can interpret and act upon, bridging the gap between conversational AI and actionable outcomes. This evolution makes LLMs not just conversational agents but problem-solving tools for real-time scenarios. What’s most exciting is the potential to build smarter, more interactive applications across industries.

6

2 Comments

🌟 Excited to share insights on the latest advancements in Natural Language Processing (NLP), Gen AI, and the transformative impact of state-of-the-art models! 🌟 🔍 In today's dynamic landscape, NLP models like GPT, BERT, Transformer-XL, XLNet, and T5 are driving unprecedented innovation across diverse domains. Each model brings unique capabilities and strengths to the table, revolutionizing how we understand, generate, and interact with text data. 🚀 Let's delve into the key differences and applications of these remarkable models: GPT (Generative Pre-trained Transformer): Known for its proficiency in generating coherent and contextually relevant text, GPT excels in tasks like text completion and story generation, making it a go-to choice for creative applications. BERT (Bidirectional Encoder Representations from Transformers): With its bidirectional context understanding, BERT is indispensable for tasks requiring nuanced understanding of sentence context, such as sentiment analysis, question answering, and named entity recognition. Transformer-XL: Overcoming the limitation of fixed-length context windows, Transformer-XL enables deeper understanding of longer-range dependencies, making it ideal for tasks involving long-document understanding and language modeling. XLNet: Adopting a permutation-based training objective, XLNet captures bidirectional context effectively without relying on masking during training, empowering tasks like text generation, machine translation, and language understanding. T5 (Text-To-Text Transfer Transformer): Revolutionizing multitask learning, T5 unifies different NLP tasks into a single text-to-text format, offering flexibility and generalization across tasks like summarization, translation, question answering, and more. 🎓 As we embrace these cutting-edge models, let's leverage their capabilities to drive innovation, solve complex challenges, and shape the future of NLP. Continuous learning and experimentation are key to unlocking the full potential of these transformative technologies. 🌐 Let's connect to explore opportunities for collaboration, knowledge sharing, and leveraging NLP advancements to drive impact in our respective domains! Together, we can harness the power of NLP to revolutionize industries and transform the way we interact with language. 🌐 #NLP #NaturalLanguageProcessing #GPT #BERT #TransformerXL #XLNet #T5 #Innovation #ArtificialIntelligence #TransformingTheFuture

4

1 Comment

🎥 Just tested out Snowflake's new Cortex AI Analyst and pitted it against my own LangGraph GPT-4o agent—interesting results! 👀 Three key areas where I think Cortex falls short: 1️⃣ Stateless design – No memory capabilities, you can't ask follow on questions or refine your prompt. 2️⃣ Lack of independent reasoning – It’s more reactive than proactive. 3️⃣ Metadata stuffing – Handling large data warehouses with static semantic stuffing could lead to scaling and maintenance issues. 💡 But I was impressed by Cortex's ability to transform vague, ambiguous questions into something concrete, tailored to the data. My take: Cortex could thrive in a framework like LangGraph. It would solve the memory issue, allow more flexible metadata handling, and function better as a query-checking or ambiguity-resolution node, potentially in parallel with GPT. I imagine Cortex is fine-tuned for Snowflake, which means it might not play as well outside of it—but if your data lives in Snowflake, integrating Cortex within LangGraph could be a powerful combination! What are your thoughts? #AI #Cortex #Snowflake #LangGraph #OpenAI #SQL

13

3 Comments

Food is more than just nourishment; it's deeply intertwined with our health, lifestyle, and culture. In the age of AI, eCommerce search is evolving from a tool into a source of inspiration. It's no longer just about showing you what you're looking for—it also offers delightful surprises you didn't know you wanted. Often, we don't know exactly what to add to our cart. Maybe we have a vague idea, a theme, or an impulse—craving traditional Asian cuisine, trying making sushi at home but needing guidance on the ingredients, planning a fun football night, aiming for a healthier lifestyle with gluten-free meals, or simply wondering, "What's for dinner?" Thanks to LLMs and their vast knowledge of brands, cuisines, recipes, nutrition, cultures, and more, we're making these experiences possible. It takes a dedicated team to seamlessly blend the world knowledge of LLMs with Instacart's insight into our users and product catalog. This involves careful engineering, thoughtful product design, and rigorous A/B testing with real-world traffic. It's been an exciting journey collaborating with such a talented team to supercharge eCommerce search with LLMs.

6

Ran a quick test on the new #OpenAI #o1mini to solve IIT-JEE Advanced 2024 maths questions. It accurately solved 13 out of 17 questions with an impressive accuracy of ~76%. Quick thoughts: 1. The hidden chain of thought is quite detailed and extensive. Breaks the problem, solves and verifies and iterates multiple times. This might increase the output token cost by a big margin. 2. Seems there are tools integrated for calculations. Much needed as #gpt4o does a bad job at it. 3. Latency is quite high. Some questions took more than 50 seconds to get answered. This needs to come down for production user-facing use cases. 4. Hallucination still exists. Check the 1st question in the link, it had the correct step-by-step reasoning but hallucinated at generating the final answer. 5. Overdoing self-thought. Saw a couple of instances where it did generate correct answers in the process but eventually made wrong answers. Would give a clearer picture once the rate limits are increased and benchmarking is done on a bigger sample. Let me know any interesting findings and thoughts 😀 https://lnkd.in/giSCezyG #GenAI #OpenAI #PhysicsWallah

139

5 Comments

✨ Hiring Alert! Level Up Your Career Game! ✨ Slide into one of the coolest tech roles at Amazon Business! We're not just building any AI - we're literally changing how the whole business world shops online! 🎮 Think of it as: • Being the main character in your own tech adventure • Unlocking new achievements with every AI solution you create • Playing in a multiplayer mode with awesome team members • Building your skill tree in cutting-edge ML/AI 💫 Your Vibe: You're that person who gets hyped about AI breakthroughs You love to experiment and aren't afraid to fail fast and learn faster Bonus points if you can turn complex data into pure magic ✨ 🔥 Why It's Fire: • Work on stuff that actually matters • Flex your AI skills on a trillion-dollar playground #AILife #FutureOfWork #ThisCouldBeYou #AmazonBusiness #CareerGoals

65

3 Comments

Resharing it: This is a great summary shared by the LinkedIn engineering team about their #llmops and LLM deployment work and their key learnings. #llms #chatgpt4 #gemini #aiml #mlops #llmops https://lnkd.in/d2bH5GRR

2

Great to see our expectations confirmed in practice! 🚀 Our Pareto front approximation method from #RecSys2024 proves its versatility beyond neural networks, now successfully applied to decision-tree-based models like LightGBM!

21

Ever wondered how random numbers work in NumPy and why setting a random seed is important? 🌟 Check out my latest blog to learn more about true vs. pseudo-randomness and the role of random seeds. #Randomness #PseudoRandomness #NumPy #PythonProgramming #Algorithms #Computing #Cryptography #ScientificComputing

21

6 Comments