Tensoic

In the world of generative AI, India isn't just keeping up with the US and China—it's leading the charge! A year ago, the scene was barren with barely a startup in sight. Now? We’ve got around 10 startups diving deep into generative AI research and over 60 whipping up cool products and wrappers from existing solutions. Welcome to the AI party, India style! Curious how India is shaking up the AI world? Read more - https://lnkd.in/gfnxj2PG Sarvam AI KissanAI Soket AI Labs Subtl.ai KOGO Tensoic CognitiveLab SML CoRover Krutrim TWO.AI Stition, Inc. NVIDIA #generativeai #indianai #techinnovation #aistartups #aiboom

Unconventional AI Learning: Join the AI Un+Conference We're thrilled to unveil the incredible lineup of speakers gracing the AI Un+Conference stage! Get ready to be inspired by leading experts who are pushing the boundaries of AI Adarsh Shirawalmath Founder, Tensoic Abhishek Upperwal CEO/Founder, Soket AI Labs Ravi Theja Desetty Developer advocate engineer, LlamaIndex This is just the first wave! We'll be dropping more speaker profiles soon, so stay connected. 📅Date: Sunday, April 28, 2024 🕐Time: 10:00 AM - 3:00 PM 🔗Register: https://lnkd.in/gjsfaQCy Here's what you can expect: * Talks & Panel Discussions: Start your day with thought-provoking discussions on the current state and future potential of AI in education, led by industry experts. * Tech Talks & Showcases: Dive deeper into specific AI advancements through a series of parallel sessions featuring 30-minute presentations by leading practitioners. * Networking Lunch: Connect with fellow AI enthusiasts, exchange ideas, and build valuable connections during the lunch break. Headstart Network Foundation IEEE AICamp Indian Institute of Management Bangalore The Product Folks #AIUnconference #AIinEducation #MachineLearning #ArtificialIntelligence #SpeakerAnnouncement #FutureofEducation #EducationTechnology

Cerule - A Tiny Mighty Vision Model Based on Google's Gemma + SigLIP we release an exciting class of vision models. Multimodal + Multilingual soon? https://lnkd.in/gFFAfsXz

Data is the new oil🛢 🚨🚀So today, we release "Guftagoo" a Hindi + Hinglish multi-turn conversational dataset comprised of 16k high quality examples Guftagoo🕉 contains multi-turn conversations on multiple topics usually revolving around daily real-life experiences. A small amount of reasoning tasks have also been added (specifically COT style reasoning and coding) with about 1k samples from Openhermes 2.5🔱 Link: https://lnkd.in/g-c46hU7 Author: Adithya Kamath #llms #indic #hindi #hinglish #data #ai

Google's Brand new model - Gemma dropped yesterday! And today we have a full fine-tune of Gemma 2B on Sarvam AI's Samvaad-hi-v1🚀🚀 https://lnkd.in/gexmffJz

💥 A small sample of what's possible in the Indic Space - Bhandara📜 A small yet textbook📘quality pre training dataset for Hindi! https://lnkd.in/gtrdiCF4 #opensource #hindi #llm #ai #indic

🤖 Weekly AI Myth Busters Myth: Closed-source AI models like OpenAI's GPT-4 and Google's Gemini are inherently superior to open-source alternatives from Meta, Hugging Face, and others. Truth: Things are changing in the AI world. While high-profile models like GPT-4 Turbo have been making headlines, the open-source community is quietly making significant strides. Take the example of Yi-34B from China's 01.AI. This open-source model is making waves by outperforming its peers in key benchmarks. On the MMLU test, which measures knowledge across 57 subjects, Yi-34B scored an impressive 76.3%, surpassing other notable open-source models like Meta's LLaMA2-70B and TII's Falcon-180B. 🚀 Why It Matters? Open-source language models offer wider accessibility and foster innovation, inviting contributions from diverse groups including those with limited resources. This approach is democratizing AI development and leading to more creative solutions. A Notable Example: Consider the release of Kan-Llama by India's Tensoic, a model adapted for the Kannada language. Such innovations are making AI more relevant and useful to different communities worldwide. Read: Should Large Language Models be Open-Sourced? Stepping into the Biggest Debates: https://lnkd.in/exuzvg6b Want to stay updated with what's happening in the Generative AI's universe? Subscribe to our LinkedIn newsletter to get weekly guides: https://lnkd.in/dCNQMGtB #llm #ai #opensource #llama2

WE HAVE A PLAYGROUND FOR Kan-LLaMA [ಕನ್-LLama]! 🚀 https://lnkd.in/gA_sRtjZ 😎 A big shout out to the folks at E2E Networks Limited for hooking us up with A100 GPUs!! They are the Biggest Providers of Cloud GPUs in India and their amazing team who guided us and Xylem AI for their amazing Superfast⚡Inference Stack with Mind Blowing Throughputs!! Do check it out!! #aicommunity #llm #india #cloud #inference #kannada

🚀 We release Kan-LLaMA [ಕನ್-LLama] — A 7B Llama-2 model, LoRA PreTrained and FineTuned on "Kannada" tokens🚀 One of the most powerful OSS LLMs can now speak Kannada! Problem? 🔴 One of the most sought out OSS LLMs — Meta's Llama-2 suffers a severe flaw. It was only trained on English tokens ! 🟡 This makes it inherently bad at generating any other language apart from english. To fix this, we've develop and release A LoRA pre-trained and fine-tuned version of Llama-2 to expand its capabilities to Kannada. 🚀🚀 We expand Llama-2's existing linguistic capabilities for Low Resource Indic languages and specifically Kannada by fine tuning on 600 Million Kannada tokens and subsequently fine-tune on SOTA Instruction Datasets. Read the blog & test out the models today! Paper and code dropping soon! Blog: https://lnkd.in/giUnpWhJ Models and Datasets: https://lnkd.in/gp_Xu-kb Contributors: Adarsh Shirawalmath, Adithya Kamath, Bharat Shetty Barkur & Raghav Ravishankar (alphabetical) #opensource #llms #kannada #multilingual #llama2

🚨We release Tiny Llama 2.5T and 3T checkpoints fine tuned on @Teknium1's Openhermes dataset🚨 "The TinyLlama project aims to pretrain a 1.1B Llama model on 3 trillion tokens. With some proper optimization, we can achieve this within a span of "just" 90 days using 16 A100-40G GPUs 🚀🚀. The training has started on 2023-09-01." . . . . https://lnkd.in/g8ibruxq https://lnkd.in/gNZq6epK