Back

Subhajit Mandal

Software Developer • 2m

1. LoRA on a reasonably small open model (best balance for small compute): apply low-rank adapters (PEFT/LoRA). Requires less GPU memory and works well for 700–3000 rows. 2. Full fine-tune (costly / heavy): only if you have >A100 GPU or cloud paid GPU. Not recommended for early MVP. 3. No-fine-tune alternative (fast & free): use retrieval + prompting (RAG) — keep base LLM and add context from your 3k+ rows. Great when compute is limited.

Reply

More like this

Recommendations from Medial

Image Description

Sandeep Prasad

Business Coach • 2m

🔥 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

See More
Reply
2
1
Image Description
Image Description

Narendra

Willing to contribut... • 1m

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

AI Engineer

AI Deep Explorer | f... • 7m

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

Swamy Gadila

Founder of Friday AI • 5m

🚨 Open AI is an Wrapper👀🤯 Hot take, but let’s break it down logically: OpenAI is not a full-stack AI company — it’s a high-level wrapper over Azure and NVIDIA. Here’s why that matters 👇 🔹 1. Infra Backbone = Microsoft Azure Almost 90%+ of Op

See More
Reply
2
4

Rahul Agarwal

Founder | Agentic AI... • 9d

Steps to building real-world AI systems. I've given a simple detailed explanation below. 𝗦𝘁𝗲𝗽 1 – 𝗗𝗲𝗽𝗹𝗼𝘆𝗺𝗲𝗻𝘁 & 𝗖𝗼𝗺𝗽𝘂𝘁𝗲 𝗟𝗮𝘆𝗲𝗿 • This is where all the 𝗵𝗲𝗮𝘃𝘆 𝗽𝗿𝗼𝗰𝗲𝘀𝘀𝗶𝗻𝗴 𝗵𝗮𝗽𝗽𝗲𝗻𝘀. • It provides the 𝗵𝗮𝗿�

See More
Reply
1
1

Rahul Agarwal

Founder | Agentic AI... • 8d

SLM vs LLM — which AI model is best for you? I’ve explained both in simple steps below. 𝗦𝗟𝗠 (𝗦𝗺𝗮𝗹𝗹 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹) (𝘴𝘵𝘦𝘱-𝘣𝘺-𝘴𝘵𝘦𝘱) Lightweight AI models built for speed, focus, and on-device execution. 1. 𝗗𝗲𝗳𝗶𝗻𝗲

See More
Reply
1
12

AI Engineer

AI Deep Explorer | f... • 8m

"A Survey on Post-Training of Large Language Models" This paper systematically categorizes post-training into five major paradigms: 1. Fine-Tuning 2. Alignment 3. Reasoning Enhancement 4. Efficiency Optimization 5. Integration & Adaptation 1️⃣ Fin

See More
Reply
1
8
Image Description

Soumya

Developer • 1y

💡An Idea to Change the Game for AI Startups: Making AI Processing Faster, Cheaper, and Effortless Running AI models like ChatGPT, DALL·E, or AlphaCode is a computing monster—they need massive power to function, which makes them expensive to operate

See More
2 Replies
4
Image Description

Animesh Kumar Singh

Hey I am on Medial • 4m

The "AI Personality Forge: Product :"Cognate" Cognate is not an AI assistant. It's a platform on your phone or computer that lets you forge, train, and deploy dozens of tiny, specialized AI "personalities" for specific tasks in seconds. How it wo

See More
1 Reply
2

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