Arnav Chavan

@adityastomar_ @nvidia @NVIDIAAI Lets catch-up sometime!

Excited to begin my summer research internship at @nvidia today. I’ll be working in the Applied Deep Learning Research team in the Santa Clara HQ office. Let me know if you are around and would like to meet!

🚀 Organizing the Efficient Qwen Competition @icmlconf ! Goal: Minimize LLM inference latency for a single GPU without breaking model quality. Prizes: $3K / $2K / $1K + present at ICML 2026, Seoul Getting Started - adaptfm.gitlab.io/call-for-compe… Leaderboard - d1krc5fcnf73gi.cloudfront.net

📣 We're extending the AdaptFM @ ICML'26 paper submission deadline to May 8! This will align better with the ICML acceptance announcements and NeurIPS deadlines. 👉 Submit your work at: openreview.net/group?id=ICML.… 🌐 CfP: adaptfm.gitlab.io

🚀 We're launching the Efficient Qwen Competition @ AdaptFM #ICML2026! How fast can you make Qwen3.5-4B run on a single A10G without breaking it? $6,000 in prizes (credits @AmazonLab126)! Submissions open May 4. 📷 adaptfm.gitlab.io/call-for-compe… #LLM #EfficientML #ModelOptimization

📢 Announcing the AdaptFM Workshop @icmlconf As foundation models grow in scale and ubiquity, the ability to adapt inference dynamically to the task and available resources becomes critical. Submission deadline: May 1, 2026 (AoE) 📍Seoul, South Korea 🌐adaptfm.gitlab.io

🔍 From tedious to streamlined! Nyun Zero's 'Nyun Kompress' module transforms model optimization with automated compression techniques that maintain performance. 🚀📊 #AIModel #TechNews #Efficiency Learn more at nyunai.com

Say goodbye to the manual grind! 🛠️ Nyun Zero automates AI model training and optimization, slashing development time and enhancing accuracy.🎯 #ModelOptimization #DeepLearning #AI #GenerativeAI #DataScience #Innovation #NyunZero Join the revolution at nyunai.com

🚀 Speeding up AI with Nyun Zero! Traditional model development takes months; we do it in days. No more long waits from concept to deployment. 🕒✨ #ModelOptimization #DeepLearning #AI #GenerativeAI #DataScience #Innovation #NyunZero 🔗 Discover how at nyunai.com

Our Team at @Nyun_AI_ recently compiled an experimental survey with a thorough codebase - Faster and Lighter LLMs: A Survey on Current Challenges and Way Forward. We provide a comprehensive overview of multiple LLM compression methodologies along with system-level approaches.

Say hello to Nyun Zero 💡- where AI meets efficiency! Reduce inference costs, speed up training, and secure your data like never before. 🛡️ Join the revolution in AI productivity. #EfficientAI #DeepLearning 🚀 Sign up here ➡️ [forms.office.com/r/NxYwkmGypG]

Rethinking Compression: Reduced Order Modelling of Latent Features in Large Language Models paper page: huggingface.co/papers/2312.07… Due to the substantial scale of Large Language Models (LLMs), the direct application of conventional compression methodologies proves impractical. The computational demands associated with even minimal gradient updates present challenges, particularly on consumer-grade hardware. This paper introduces an innovative approach for the parametric and practical compression of LLMs based on reduced order modelling, which entails low-rank decomposition within the feature space and re-parameterization in the weight space. Notably, this compression technique operates in a layer-wise manner, obviating the need for a GPU device and enabling the compression of billion-scale models within stringent constraints of both memory and time. Our method represents a significant advancement in model compression by leveraging matrix decomposition, demonstrating superior efficacy compared to the prevailing state-of-the-art structured pruning method.

@chenlin_meng @CVPR Any plans to release the code @chenlin_meng ?

We will be presenting our @CVPR Award Candidate paper On Distillation of Guided Diffusion Models tomorrow Wed 21 Jun 4:30-6 p.m. at West Building Exhibit Halls ABC 186! Super honored that our paper is selected as one of the 12 award candidates! 🙏 #CVPR2023

@_akhaliq Thanks for your tweet. Code - github.com/Arnav0400/ViT-… Website - sites.google.com/view/generaliz…

One-for-All: Generalized LoRA for Parameter-Efficient Fine-tuning paper page: huggingface.co/papers/2306.07… present Generalized LoRA (GLoRA), an advanced approach for universal parameter-efficient fine-tuning tasks. Enhancing Low-Rank Adaptation (LoRA), GLoRA employs a generalized prompt module to optimize pre-trained model weights and adjust intermediate activations, providing more flexibility and capability across diverse tasks and datasets. Moreover, GLoRA facilitates efficient parameter adaptation by employing a scalable, modular, layer-wise structure search that learns individual adapter of each layer. Originating from a unified mathematical formulation, GLoRA exhibits strong transfer learning, few-shot learning and domain generalization abilities, as it adjusts to new tasks through additional dimensions on weights and activations. Comprehensive experiments demonstrate that GLoRA outperforms all previous methods in natural, specialized, and structured benchmarks, achieving superior accuracy with fewer parameters and computations on various datasets. Furthermore, our structural re-parameterization design ensures that GLoRA incurs no extra inference cost, rendering it a practical solution for resource-limited applications.

@jquave huggingface.co/papers/2306.07…

If your working on any of these, DM me: 1. In memory vector database 2. Generalized LoRA adapters dev environment+marketplace. (npm for LoRA adapters) 3. Marketplace for ongoing rlhf tasks 4. LLM schema fine tuning for consistent outputs in, for example, YAML