Back

Sandeep Prasad

Business Coach • 5m

🔥 Government set to name ~8 Indian teams for foundational model incentives next week – second-round beneficiaries may include BharatGen; GPU access remains tight as only ~17,374 of planned 34,333 GPUs are installed so far. 🤔 Why It Matters – More subsidised compute means faster India‑tuned models, but the GPU crunch could slow training unless procurement accelerates or inference‑efficient approaches are prioritised. 🚀 Action/Example – Founders should prepare grant docs and pivot to efficient training/inference (LoRA, distillation, 4‑bit quant) to ride the incentive window despite supply constraints. 🎯 Who Benefits – AI researchers, Indic LLM builders, and startups focused on low‑cost inference at scale. Tap ❤️ if you like this post.

Reply
5
1

More like this

Recommendations from Medial

Aditya Karnam

Hey I am on Medial • 11m

"Just fine-tuned LLaMA 3.2 using Apple's MLX framework and it was a breeze! The speed and simplicity were unmatched. Here's the LoRA command I used to kick off training: ``` python lora.py \ --train \ --model 'mistralai/Mistral-7B-Instruct-v0.2' \ -

See More
Reply
1

Inactive

AprameyaAI • 1y

Meta has released Llama 3.1, the first frontier-level open source AI model, with features such as expanded context length to 128K, support for eight languages, and the introduction of Llama 3.1 405B. The model offers flexibility and control, enabli

See More
Reply
2
9
Image Description
Image Description

Narendra

Willing to contribut... • 3m

I fine-tuned 3 models this week to understand why people fail. Used LLaMA-2-7B, Mistral-7B, and Phi-2. Different datasets. Different methods (full tuning vs LoRA vs QLoRA). Here's what I learned that nobody talks about: 1. Data quality > Data quan

See More
2 Replies
10
1

PraYash

Technology, Business... • 4m

When AI changed the rules, cloud computing had to change too. And that’s exactly where Oracle took the lead. Most cloud giants like AWS, Azure, and GCP still rely on virtualization — where resources like CPU, GPU, and memory are shared across users.

See More
Reply
6

AI Engineer

AI Deep Explorer | f... • 10m

LLM Post-Training: A Deep Dive into Reasoning LLMs This survey paper provides an in-depth examination of post-training methodologies in Large Language Models (LLMs) focusing on improving reasoning capabilities. While LLMs achieve strong performance

See More
Reply
2
Image Description

Rahul Agarwal

Founder | Agentic AI... • 2m

4 different ways of training LLM's. I've given a simple detailed explanation below. 1.) 𝗔𝗰𝗰𝘂𝗿𝗮𝘁𝗲 𝗗𝗮𝘁𝗮 𝗖𝘂𝗿𝗮𝘁𝗶𝗼𝗻 (𝘀𝘁𝗲𝗽-𝗯𝘆-𝘀𝘁𝗲𝗽) Prepares clean, consistent, and useful data so the model learns effectively. 1. Collect text

See More
Reply
1
9
1

Vansh Khandelwal

Full Stack Web Devel... • 15d

Small language models (SLMs) are compact NLP models optimized for edge devices (smartphones, IoT, embedded systems). With fewer parameters they enable faster, low‑latency on‑device inference, improving privacy by avoiding cloud transfers, reducing en

See More
Reply
1
5
Image Description
Image Description

Mada Dhivakar

Let’s connect and bu... • 8m

Why Grok AI Outperformed ChatGPT & Gemini — Without Spending Billions In 2025, leading AI companies invested heavily in R&D: ChatGPT: $75B Gemini: $80B Meta: $65B Grok AI, developed by Elon Musk's xAI, raised just $10B yet topped global benchmar

See More
1 Reply
7
1

Varun Bhambhani

 • 

Medial • 11d

🗞️ Medial Bulletin 🤖 AI & Automation • OpenAI launches Agent Tasks for enterprises, enabling AI agents to run multi-step workflows with approval checkpoints. • Salesforce extends Einstein Copilot to CRM analytics, adding AI-based forecasting and

See More
Reply
3
9
Image Description

Parampreet Singh

Python Developer 💻 ... • 12m

3B LLM outperforms 405B LLM 🤯 Similarly, a 7B LLM outperforms OpenAI o1 & DeepSeek-R1 🤯 🤯 LLM: llama 3 Datasets: MATH-500 & AIME-2024 This has done on research with compute optimal Test-Time Scaling (TTS). Recently, OpenAI o1 shows that Test-

See More
1 Reply
5

Download the medial app to read full posts, comements and news.