Junwoo Yun

SQL to Pandas

2

A platform to learn and gain insights as per interest. Look into the GitHub Repositories to gain knowledge on Machine Learning and Artificial Intelligence.

2

I recently saw an amazing video on 𝐆𝐏𝐔 𝐨𝐩𝐭𝐢𝐦𝐢𝐳𝐚𝐭𝐢𝐨𝐧 𝐰𝐨𝐫𝐤𝐬𝐡𝐨𝐩 organized by Chip Huyen. Here is the video link: https://lnkd.in/gVm-KQtt If you are an ML Engineer my 2 cents to improving LLM deployments on GPUs is to use 𝐐𝐮𝐚𝐧𝐭𝐢𝐳𝐚𝐭𝐢𝐨𝐧 and 𝐨𝐩𝐭𝐢𝐦𝐢𝐳𝐞 𝐊𝐕-𝐜𝐚𝐜𝐡𝐢𝐧𝐠 as mentioned in the video. So how can we go about doing it? Here I provide a few open source projects which you can use today to heavily optimize your LLM deployments across both servers and on device. I am pretty sure (depending on your server load and number of users) you are looking at saving potentially millions of dollars! 🎁 𝐐𝐮𝐚𝐧𝐭𝐢𝐳𝐚𝐭𝐢𝐨𝐧 If you are deploying models on servers such as GCP/AWS/Azure where you are generally batching the requests then you are better off with using SmoothQuant(https://lnkd.in/gDxt4F_Z) to quantize the models to 8𝐛𝐢𝐭𝐬 𝐰𝐞𝐢𝐠𝐡𝐭𝐬 𝐚𝐧𝐝 8𝐛𝐢𝐭𝐬 𝐚𝐜𝐭𝐢𝐯𝐚𝐭𝐢𝐨𝐧𝐬. If you are deploying models on device (Mobile, Laptop) etc. Then you should use AWQ(https://lnkd.in/gGNhPcPR) to quantize the models to 4𝐛𝐢𝐭𝐬 𝐰𝐞𝐢𝐠𝐡𝐭𝐬 𝐚𝐧𝐝 16𝐛𝐢𝐭𝐬 𝐚𝐜𝐭𝐢𝐯𝐚𝐭𝐢𝐨𝐧𝐬. 𝐈𝐧 𝐛𝐨𝐭𝐡 𝐨𝐟 𝐭𝐡𝐞𝐬𝐞 𝐜𝐚𝐬𝐞𝐬 𝐲𝐨𝐮 𝐰𝐢𝐥𝐥 𝐞𝐧𝐣𝐨𝐲 𝐚𝐜𝐜𝐮𝐫𝐚𝐜𝐲 𝐰𝐢𝐭𝐡𝐢𝐧 1% 𝐨𝐟 𝐭𝐡𝐞 𝐨𝐫𝐢𝐠𝐢𝐧𝐚𝐥 𝐁𝐅𝐥𝐨𝐚𝐭16 model. 🎁 𝐏𝐚𝐠𝐞𝐝 𝐊𝐕 𝐂𝐚𝐜𝐡𝐢𝐧𝐠 𝐚𝐧𝐝 𝐊𝐕 𝐜𝐚𝐜𝐡𝐞 𝐫𝐞𝐮𝐬𝐞. Paged KV caching is a way to use non-contiguous blocks of memory to store KV caches so that we aren't wasting precious GPU memory. KV cache reuse is: if two sets of queries are similar, we can reuse the same KV caches instead of recaching them. Both of these might sound intimidating to implement but, the great news is that, this is all part of vLLM(https://lnkd.in/g7uyGPQA) which is based on PagedAttention(https://lnkd.in/gRExM3c5). So in short just use the vLLM as your deployment server! #machinelearning #deployment #inference #llm #chatgpt #inference #costsavings #artificialintelligence

147

댓글 2

Bookmarking this will definitely help you in the future!

1

Start your #AI journey now.

7

**TEST V0001OSS MODEL BASE ROLLOUT.: ***MY OWN COLAB _or_ COLABORATORY (imported .py package module scripts-> INFO-HIDE **HINT-HINT** CLEAR AS DAY -> PRIOR 4 DEV (PURE LIN.//WIN.) (SYS. TEST ERRORS CAUGHT IN MY OWN VIRTUAL SYSTEM (ISO. BUBBLE)) -> JSON VERSIONING/MESSAGE PASSING MY IP FULLY ROLL-OUT FILES (RD//BLU//GRN TEAMs ARROWs; COPY FILES) VERSIONING -> ***DRY-RUN BOX ONLY TEST CASE(S) -> ONLY HYBRID OOP NO UNAUTHORIZED [[KEY-WORDS SENSITIVE AI-PROMPTS [[AUTO-AI, REVERSE-ENGINEERING (-OR- REVENG) COMBSEC KEY 🚩]] AUTO-EXECUTION PROGRAM(‘s) OOP ANALYTICS MODULAR PORTABLE OOP HYBRID CLASS (ran as .py or .ipynb out-of-the-box downloaded directly from colab or colaboratory [[YOUR FIRM FAVORITE ENV. + ((REQs||DEPs)) and pip freeze method]] for generated requirements.txt files per .py file(s), imported .py packages reqs/deps (requirements/dependencies) HOUSING MY ABL(‘s) FUNC(s) (or my own REVENG MY IP PROCESS ROOT SYS OVER USER funct. funcs. (or functions)) SINGLE SHARABLE sub-repo LINK(‘S)) -> BASE UNIT-TEST CASE v0001 ex.: sample: **YOUR AUTO-DEMO OUTPUTS** -> ****BIGGEST RED FLAGS -> YOUR FIRM COMBSEC KEY ROLL-OUT; YOUR SELECTED SRVCS. *****SINGLE SHARABLE LINK -> SHARED SINGLE ACCESS-POINT(S) OF RISK -> YOUR FIRM RE-ASSESS YOUR Sys-dot RISK(s)

1

💥💥💥 OpenCoder: The Open Cookbook for Top-Tier Code Large Language Models Siming Huang, Tianhao Cheng, Jason Klein Liu, Jiaran Hao, Liuyihan Song, Yang Xu, J. Yang, J.H. Liu, Chenchen Zhang, Linzheng Chai, Ruifeng Yuan, Zhaoxiang Zhang, Jie Fu, Qian Liu, Ge Zhang, Zili Wang, Yuan Qi, Yinghui Xu, Wei Chu Abstract Large language models (LLMs) for code have become indispensable in various domains, including code generation, reasoning tasks and agent this http URL open-access code LLMs are increasingly approaching the performance levels of proprietary models, high-quality code LLMs suitable for rigorous scientific investigation, particularly those with reproducible data processing pipelines and transparent training protocols, remain limited. The scarcity is due to various challenges, including resource constraints, ethical considerations, and the competitive advantages of keeping models advanced. To address the gap, we introduce OpenCoder, a top-tier code LLM that not only achieves performance comparable to leading models but also serves as an ``open cookbook'' for the research community. Unlike most prior efforts, we release not only model weights and inference code, but also the reproducible training data, complete data processing pipeline, rigorous experimental ablation results, and detailed training protocols for open scientific research. Through this comprehensive release, we identify the key ingredients for building a top-tier code LLM: (1) code optimized heuristic rules for data cleaning and methods for data deduplication, (2) recall of text corpus related to code and (3) high-quality synthetic data in both annealing and supervised fine-tuning stages. By offering this level of openness, we aim to broaden access to all aspects of a top-tier code LLM, with OpenCoder serving as both a powerful model and an open foundation to accelerate research, and enable reproducible advancements in code AI. 👉 https://lnkd.in/dCPNQnTT #machinelearning

1

[arXiv] Deep Learning job interview questions, fully solved & covering a wide range of key AI topics https://lnkd.in/gWkziAH5

8

𝐈𝐧𝐭𝐞𝐫𝐯𝐢𝐞𝐰 𝐐1: 𝐖𝐡𝐚𝐭 𝐚𝐫𝐞 𝐭𝐡𝐞 𝐝𝐢𝐟𝐟𝐞𝐫𝐞𝐧𝐭 𝐰𝐚𝐲𝐬 𝐲𝐨𝐮 𝐜𝐚𝐧 𝐥𝐨𝐚𝐝 𝐝𝐚𝐭𝐚 𝐢𝐧𝐭𝐨 𝐑𝐞𝐝𝐬𝐡𝐢𝐟𝐭? 𝑵𝒆𝒘 𝑩𝒂𝒕𝒄𝒉 𝒕𝒐 𝒔𝒕𝒂𝒓𝒕 𝑱𝒖𝒍𝒚 28𝒕𝒉- 𝐉𝐨𝐢𝐧 𝐚𝐥𝐨𝐧𝐠 𝐰𝐢𝐭𝐡 47 𝐚𝐦𝐚𝐳𝐢𝐧𝐠 𝐩𝐞𝐨𝐩𝐥𝐞 𝐰𝐡𝐨 𝐡𝐚𝐯𝐞 𝐞𝐧𝐫𝐨𝐥𝐥𝐞𝐝. 1 ) From Data Lake S3 to Redshift table using Redshift spectrum 2) Using S3 Copy to load data from S3 into Redshift table utilizing the massive parallel processing capability of Redshift 3) Use S3 Auto Copy - Set up this job in Redshift and everytime data is added in Data Lake , it gets loaded to Redshift 4) Use Glue ETL to load data into Redshift 5) Use EMR to load data into Redshift 𝐈𝐧𝐭𝐞𝐫𝐯𝐢𝐞𝐰 𝐐2: 𝐇𝐨𝐰 𝐝𝐨 𝐲𝐨𝐮 𝐠𝐞𝐭 𝐝𝐚𝐭𝐚 𝐨𝐮𝐭 𝐨𝐟 𝐑𝐞𝐝𝐬𝐡𝐢𝐟𝐭? 1) Use UNLOAD command to unload data into S3 Datalake #aws #dataengineering ✅ Enrollments for Version 3 of 𝐬𝐜𝐞𝐧𝐚𝐫𝐢𝐨 𝐛𝐚𝐬𝐞𝐝 𝐬𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞𝐝 𝐡𝐚𝐧𝐝𝐬-𝐨𝐧 𝐑𝐞𝐚𝐥-world AWS Data Engineering ( RADE™) program is ON ! 𝑵𝒆𝒘 𝑩𝒂𝒕𝒄𝒉 𝒕𝒐 𝒔𝒕𝒂𝒓𝒕 𝑱𝒖𝒍𝒚 28𝒕𝒉- 𝐉𝐨𝐢𝐧 𝐚𝐥𝐨𝐧𝐠 𝐰𝐢𝐭𝐡 47 𝐚𝐦𝐚𝐳𝐢𝐧𝐠 𝐩𝐞𝐨𝐩𝐥𝐞 𝐰𝐡𝐨 𝐡𝐚𝐯𝐞 𝐞𝐧𝐫𝐨𝐥𝐥𝐞𝐝. ✅𝐈𝐟 𝐲𝐨𝐮 𝐚𝐫𝐞 𝐚𝐜𝐭𝐢𝐨𝐧-𝐨𝐫𝐢𝐞𝐧𝐭𝐞𝐝, fill the form below to be get invited for the webinar for RADE™ program V3

110

댓글 5

💥💥💥 mDPO: Conditional Preference Optimization for Multimodal Large Language Models Fei Wang1, Wenxuan Zhou1, James Y. Huang1, Nan Xu1  Sheng Zhang3, Hoifung Poon3, Muhao Chen2 | 1USC, 2UC Davis, 3Microsoft Research 𝐓𝐡𝐞 𝐏𝐢𝐭𝐟𝐚𝐥𝐥: 𝐔𝐧𝐜𝐨𝐧𝐝𝐢𝐭𝐢𝐨𝐧𝐚𝐥 𝐏𝐫𝐞𝐟𝐞𝐫𝐞𝐧𝐜𝐞 Through controlled experiments, we find that MLLMs can perform similarly even if all images are removed from DPO training. While DPO aims to compute implicit rewards conditioned on all input modalities, it may prioritize language-only preferences and overlook the image condition, leading to hallucinations. 𝐒𝐨𝐥𝐮𝐭𝐢𝐨𝐧: ᴍ𝐃𝐏𝐎 ᴍDPO, a multimodal DPO objective that jointly optimizes text and image preferences, with absolute reward anchors to address problems caused by relative preferences. 𝐊𝐞𝐲 𝐓𝐚𝐤𝐞𝐚𝐰𝐚𝐲𝐬 • Unconditional preference is common and harmful in DPO for multimodal scenarios, where the model often overlooks visual signals. • It is essential to jointly optimize preferences across different modalities. • ᴍDPO effectively addresses the unconditional preference issue and significantly improves model performance, particularly in reducing hallucinations. 👉 https://lnkd.in/dQDxYwmJ #machinelearning

1

All great resources but really like applied ml

4

𝗘𝘅𝗰𝗶𝘁𝗶𝗻𝗴 𝗡𝗲𝘄𝘀 𝘄𝗶𝘁𝗵𝗶𝗻 𝘁𝗵𝗲 𝗟𝗮𝗿𝗴𝗲 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 (𝗟𝗟𝗠𝘀) 𝗦𝗽𝗮𝗰𝗲! PyTorch has officially released TorchTune, a new tool designed to fine-tune Large Language Models (LLMs) with ease and efficiency. This groundbreaking tool promises to streamline the fine-tuning process, making it more accessible to researchers and practitioners alike. https://lnkd.in/eMf6dNZJ 𝗧𝗼𝗿𝗰𝗵𝘁𝘂𝗻𝗲 𝗢𝗳𝗳𝗲𝗿𝘀: ▪ 𝗣𝘆𝗧𝗼𝗿𝗰𝗵-𝗻𝗮𝘁𝗶𝘃𝗲 𝗟𝗟𝗠 𝗶𝗺𝗽𝗹𝗲𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻𝘀: Build on familiar ground with composable building blocks for popular LLMs. ️ ▪ 𝗠𝗲𝗺𝗼𝗿𝘆-𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗳𝗶𝗻𝗲-𝘁𝘂𝗻𝗶𝗻𝗴 𝗿𝗲𝗰𝗶𝗽𝗲𝘀: Train on consumer-grade GPUs with 24GB VRAM – no beefy machines needed! ▪ 𝗘𝘅𝘁𝗲𝗻𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆 𝗮𝗻𝗱 𝗲𝗮𝘀𝗲 𝗼𝗳 𝘂𝘀𝗲:  YAML configs and support for common dataset formats streamline your workflow. 𝗙𝗼𝗿 𝘁𝗵𝗼𝘀𝗲 𝗶𝗻𝘁𝗲𝗿𝗲𝘀𝘁𝗲𝗱 𝗶𝗻 𝗱𝗶𝘃𝗶𝗻𝗴 𝗱𝗲𝗲𝗽𝗲𝗿, 𝗵𝗲𝗿𝗲 𝗮𝗿𝗲 𝘀𝗼𝗺𝗲 𝗿𝗲𝘀𝗼𝘂𝗿𝗰𝗲𝘀: Code: https://lnkd.in/gEgETe5t Blog: https://lnkd.in/gP3mctA6 Tutorials: https://lnkd.in/gvh2eEsU This is a fantastic opportunity to explore and innovate, I'd be very interested in hearing your thoughts. ▪ 𝗪𝗵𝗮𝘁’𝘀 𝘆𝗼𝘂𝗿 𝘁𝗮𝗸𝗲 𝗼𝗻 𝘁𝗵𝗶𝘀?  ▪ 𝗛𝗼𝘄 𝗱𝗼 𝘆𝗼𝘂 𝘀𝗲𝗲 𝗧𝗼𝗿𝗰𝗵𝗧𝘂𝗻𝗲 𝗶𝗺𝗽𝗮𝗰𝘁𝗶𝗻𝗴 𝘆𝗼𝘂𝗿 𝗰𝘂𝗿𝗿𝗲𝗻𝘁 𝗽𝗿𝗼𝗷𝗲𝗰𝘁𝘀 𝗼𝗿 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵? ▪ 𝗛𝗼𝘄 𝗱𝗼 𝘆𝗼𝘂 𝘁𝗵𝗶𝗻𝗸 𝗧𝗼𝗿𝗰𝗵𝗧𝘂𝗻𝗲 𝘄𝗶𝗹𝗹 𝗶𝗺𝗽𝗮𝗰𝘁 𝘁𝗵𝗲 𝗳𝘂𝘁𝘂𝗿𝗲 𝗼𝗳 𝗺𝗮𝗰𝗵𝗶𝗻𝗲 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗮𝗻𝗱 𝗔𝗜 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁? #AI #PyTorch #TorchTune #LLMs #MachineLearning

24

댓글 1

💥💥💥 DeepSeek-Prover-V1.5: Harnessing Proof Assistant Feedback for Reinforcement Learning and Monte-Carlo Tree Search Huajian Xin, Z.Z. Ren, Junxiao Song, Zhihong Shao, Wanjia Zhao, Haocheng Wang, Bo Liu, Liyue Zhang, Xuan Lu, Qiushi Du, Wenjun Gao, Qihao Zhu, Dejian Yang, Zhibin Gou, Z.F. Wu, Fuli Luo, Chong Ruan Abstract We introduce DeepSeek-Prover-V1.5, an open-source language model designed for theorem proving in Lean 4, which enhances DeepSeek-Prover-V1 by optimizing both training and inference processes. Pre-trained on DeepSeekMath-Base with specialization in formal mathematical languages, the model undergoes supervised fine-tuning using an enhanced formal theorem proving dataset derived from DeepSeek-Prover-V1. Further refinement is achieved through reinforcement learning from proof assistant feedback (RLPAF). Beyond the single-pass whole-proof generation approach of DeepSeek-Prover-V1, we propose RMaxTS, a variant of Monte-Carlo tree search that employs an intrinsic-reward-driven exploration strategy to generate diverse proof paths. DeepSeek-Prover-V1.5 demonstrates significant improvements over DeepSeek-Prover-V1, achieving new state-of-the-art results on the test set of the high school level miniF2F benchmark (63.5%) and the undergraduate level ProofNet benchmark (25.3%). 👉 https://lnkd.in/dgG9eqk7 👉 https://lnkd.in/dRvTVe3p #machinelearning

1

🚀 𝗛𝗼𝘄 “𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗬𝗲𝗮𝗿𝗻𝗶𝗻𝗴” 𝗯𝘆 Andrew Ng 𝗧𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗲𝗱 𝗠𝘆 𝗠𝗟 𝗝𝗼𝘂𝗿𝗻𝗲𝘆! 📚 I recently completed 𝗔𝗻𝗱𝗿𝗲𝘄 𝗡𝗴’𝘀 𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗬𝗲𝗮𝗿𝗻𝗶𝗻𝗴, and it has revolutionized my approach to machine learning. 🌟 𝗛𝗲𝗿𝗲’𝘀 𝗵𝗼𝘄 𝘁𝗵𝗶𝘀 𝗯𝗼𝗼𝗸 𝗵𝗮𝘀 𝗺𝗮𝗱𝗲 𝗮 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝗰𝗲: 1️⃣ 𝙎𝙩𝙧𝙪𝙘𝙩𝙪𝙧𝙞𝙣𝙜 𝙈𝙇 𝙋𝙧𝙤𝙟𝙚𝙘𝙩𝙨: Ng’s strategies for setting up development and test sets optimized my workflows. Selecting data from the correct distribution was a game-changer! 2️⃣ 𝙈𝙖𝙨𝙩𝙚𝙧𝙞𝙣𝙜 𝘽𝙞𝙖𝙨 & 𝙑𝙖𝙧𝙞𝙖𝙣𝙘𝙚: The book dives deep into diagnosing and reducing bias and variance. Ng’s clear, step-by-step techniques are invaluable for model optimization. 3️⃣ 𝑸𝒖𝒊𝒄𝒌 𝑰𝒕𝒆𝒓𝒂𝒕𝒊𝒐𝒏 & 𝑬𝒓𝒓𝒐𝒓 𝑨𝒏𝒂𝒍𝒚𝒔𝒊𝒔: Emphasizing quick system builds and iterative error analysis helped me identify and prioritize improvements effectively. 4️⃣ 𝑶𝒑𝒕𝒊𝒎𝒊𝒛𝒊𝒏𝒈 𝑴𝒆𝒕𝒓𝒊𝒄𝒔: Establishing a single-number evaluation metric aligned my team’s focus, ensuring faster and more efficient progress. 5️⃣ 𝑨𝒄𝒄𝒆𝒔𝒔𝒊𝒃𝒍𝒆 𝑰𝒏𝒔𝒊𝒈𝒉𝒕𝒔: Whether you’re an ML researcher or a practitioner, the book’s actionable advice is beneficial for real-world applications. 💡 If you’re diving into machine learning or refining your ML strategy, Machine Learning Yearning is a must-read! Andrew Ng simplifies complex topics and provides practical approaches to achieving breakthroughs. It’s a rare gem that teaches you not just how to build models, but how to think about solving ML problems. 🚀 Highly recommend it to anyone serious about machine learning! #MachineLearning #AI #AndrewNg #MLBooks #DataScience #TechInnovation IBM #BiasVariance #DeepLearning DeepLearning.AI Google DeepMind

31

I'm thrilled to share that our latest research paper, "Harnessing the Power of Hadoop and ChatGPT for Trending Product Recommendations," has been accepted for presentation at the FOCUS-2024 Conference this December! This paper is a collaborative effort, with Samuel Tan Joo Woon as the lead author and Hin Sheik Wee and myself as co-authors. Together, we explored integrating big data platforms like Hadoop with advanced AI models like ChatGPT to create innovative solutions for product recommendation systems. #Research #AI #Hadoop #ChatGPT #ProductRecommendations #FOCUS2024 #Teamwork

65

댓글 12

Data types in Redshift tables: 52.1% — varchar 10.2% — numeric(p, s) ¹ 9.1% — integer 7.0% — bigint 6.2% — timestamp w/o tz 4.5% — double 3.9% — boolean 2.2% — date 2.1% — smallint 1.7% — char(n) 0.4% — float ² 0.4% — timestamp w/ tz This sums to 99.8%. Other reported numbers: ≈0.1% — json ≈0.02% — times / intervals ≈0.02% — geography Data comes from a recent paper published by Amazon: "Why TPC Is Not Enough: An Analysis of the Amazon Redshift Fleet". varchar is a flexible data type. It is sometimes used to store data that fits one of the other types: - date as string ➜ e.g. "2024-08-02" - boolean as string ➜ e.g. "yes" and "no" - json as string - … Because varchar is used so often, it's important for Redshift to ensure efficient string storage and processing. This is probably also true for other cloud data warehouses. Link to paper in comment. ¹ also often called decimal(p, s) ² I think "float" maps to the single-precision REAL type (and not to FLOAT, which is an alias of DOUBLE PRECISION in Redshift)

17

🚀 𝐄𝐱𝐜𝐢𝐭𝐢𝐧𝐠 𝐀𝐈 & 𝐒𝐐𝐋 𝐈𝐧𝐭𝐞𝐠𝐫𝐚𝐭𝐢𝐨𝐧! I recently tried out OpenAI's GPT-4o-mini with SQL using the 𝐩𝐫𝐨𝐦𝐩𝐭() function, inspired by MotherDuck‘s innovative approach. This feature now allows us to 𝐠𝐞𝐧𝐞𝐫𝐚𝐭𝐞, 𝐬𝐮𝐦𝐦𝐚𝐫𝐢𝐳𝐞, 𝐚𝐧𝐝 𝐞𝐱𝐭𝐫𝐚𝐜𝐭 𝐬𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞𝐝 𝐝𝐚𝐭𝐚  directly from text—all within a SQL query, no extra infrastructure needed! 💡 It's truly a game-changer in simplifying data handling. 📊 💻 The 𝐩𝐫𝐨𝐦𝐩𝐭() function can even generate structured outputs using the struct and struct_descr parameters. This means you can specify the structure of typed return values, making it easier than ever to integrate AI-generated data into your analytics workflows. ✅ OpenAI's structured outputs ensure that the generated data aligns perfectly with the schema you define. Excited to keep exploring the endless possibilities of blending AI and SQL! 🔥 #AI 🤖 #SQL 💾 #OpenAI #SLM #LLM #DataScience #Automation #MotherDuck #StructuredData #Innovation https://lnkd.in/gE-aDha9

35

댓글 1