Artificial Intelligence l l AI Updates

🔗 GitHub_Link ❇️ Distributed Swarm Trajectory Optimization for Formation Flight in Dense Environments. Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ PDNet: Toward Better One-Stage Object Detection With Prediction Decoupling 🔥 #ObjectDetection Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ I2V-Adapter: A General Image-to-Video Adapter for Diffusion Models #DiffusionModels Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ Fine-tune SAM (Segment Anything Model) for computer vision tasks such as semantic segmentation, matting, detection ... in specific scenarios #SAM Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ Prompt Segment Anything 🔥 #SAM Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ USE OPEN-SOURCE LLMS IN POSTGRESQL with Ollama and new open-source extension pgai 🦙 𝗪𝗵𝗮𝘁 𝗶𝘀 𝗢𝗹𝗹𝗮𝗺𝗮? Ollama is an easy and popular way to use open-source language models like Llama 3, Mistral, Phi 3, and Gemma. Unlike proprietary models, open-source models are private, free (hardware costs aside), can run locally, and are customizable. 🐘 𝗪𝗵𝗮𝘁 𝗶𝘀 𝗽𝗴𝗮𝗶? Pgai is an open-source PostgreSQL extension that integrates AI models with PostgreSQL data, simplifying AI engineering for developers familiar with PostgreSQL and facilitating RAG and search. 🧰 𝗪𝗵𝗮𝘁 𝗰𝗮𝗻 𝗜 𝗱𝗼 𝘄𝗶𝘁𝗵 𝗽𝗴𝗮𝗶 𝗮𝗻𝗱 𝗢𝗹𝗹𝗮𝗺𝗮? Create embeddings on PostgreSQL data using models like BERT and Llama 3, storing them in pgvector for easy search and RAG. Perform RAG and LLM reasoning tasks using models like Llama 3, Mistral, and Gemma, enabling summarization, categorization, and data enrichment via SQL queries. #LLMs Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ TF-ID: Table/Figure IDentifier for academic papers. Seeing the open-source community develop small, cost-effective customized vision-language models (VLMs) that outperform the much larger closed-source APIs is really impressive. One of them is the TF-ID model by Yifei Huang. It's a fine-tuned version of Florence-2, the small but very powerful VLM by Microsoft. TF-ID (Table/Figure IDentifier) is a family of object detection models finetuned to extract tables and figures in academic papers. Interestingly, the author labeled 4600 images by hand, ensuring high data quality! #VLMs Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ ChartGemma: Visual Instruction-tuning for Chart Reasoning in the Wild It's really cool to see the open-source community creating small and cheap customized vision-language models (VLMs) that outperform the much larger closed-source APIs. ChartGemma is a fine-tuned version of PaliGemma created by Megh Thakkar and team, which excels at answering questions regarding charts and plots. The idea is pretty simple: first use a closed-source API like Gemini 1.5 Flash to collect training data, then fine-tune the open PaliGemma model on it. You end up with a model that is much smaller and cheaper to run for this specific niche task, and it outperforms the closed-source APIs! 🔥 #VLMs Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ Customize your NeRF: Adaptive Source Driven 3D Scene Editing via Local-Global Iterative Training #NeRF Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ Shape of Motion: 4D Reconstruction from a Single Video Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ DataComp for Language Models Apple has entered the game! Apple just released a 7B open-source LLM, weights, training code, and dataset! 👀 🧠 7B base model, trained on 2.5T tokens on an open datasets 🌐 Primarily English data and a 2048 context window 📈 Combined DCLM-BASELINE, StarCoder, and ProofPile2 data 🏆 MMLU 0.6372 > Mistral & < Llama3 🔓 Open License with Apple Sample Code License 📊 Matches closed-dataset models like Mistral 🔬 Trained using PyTorch with OpenLM framework 🤗 Available on Hugging Face and in Transformers #LLMs Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ DeepSolo: Let Transformer Decoder with Explicit Points Solo for Text Spotting #SceneTextSpotting Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ MetaSeg: Packaged version of the Segment Anything 🔥 #SAM Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ Robotic Transformer 2 (RT-2): The Vision-Language-Action Model #Robotics Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ HumanoidBench: Simulated Humanoid Benchmark for Whole-Body Locomotion and Manipulation #Robotics Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ OmniTokenizer: one model and one weight for image-video joint tokenization #VideoGeneration Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ Big news. Andrej Karpathy is launching a new AI Education company called Eureka labs. Their first product will be the world's best AI course, LLM101n 🔥. #LLMs Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates

🔗 GitHub_Link ❇️ PowerPaint: A Versatile Image Inpainting Model 🔆🖌 Join my channel: 👇👇👇👇👇👇 https://t.me/Artificial_Intelligence_Updates