Kicking off an 𝗔𝗰𝘁𝗶𝘃𝗮𝘁𝗶𝗼𝗻 𝗙𝘂𝗻𝗰𝘁𝗶𝗼𝗻𝘀 𝗶𝗻 𝗗𝗲𝗲𝗽 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝘀𝗲𝗿𝗶𝗲𝘀 ⚡
🔹 𝗖𝗼𝗿𝗲 𝗙𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻𝘀
Day 1: What, Why & When of Activation Functions
Day 2: ReLU — The Backbone of Deep Learning
Day 3: Sigmoid — Probabilities & Vanishing Gradients
Day 4: Tanh — Zero-Centered Nonlinearity
Day 5: Softmax — Multi-Class Classification Output
Day 6: Leaky ReLU — Fixing Dead Neurons
Day 7: ELU & SELU — Stability & Self-Normalization
🔹𝗠𝗼𝗱𝗲𝗿𝗻 𝗯𝘂𝘁 𝗦𝘁𝗮𝗻𝗱𝗮𝗿𝗱
Day 8: Swish — Smooth Alternative to ReLU
Day 9: GELU — Transformer & LLM Standard
Day 10: Mish — Accuracy-Focused Smooth Activation
Day 11: Choosing Activations for CNNs, RNNs & Transformers
Day 12: Failure Modes — Vanishing, Exploding & Dead Neurons
Day 13: Optimization View — How Activations Shape Training
Day 14: Final Practical Guide — How to Choose the Right Activation
From basic non-linearity to production-grade deep learning.
No unnecessary theory.
Only what you actually need to understand and build neural networks correctly.
Been working on IntervieX lately.
An AI mock interview platform that helps you practice smarter, not harder
Try it out and let me know 😁
https://t.co/hLyOjTUdPQ
Done with college
Last day it was,
It was fun until last moment where all became emotional suddenly 🥹
Thank you @btwitsvardhan , @adavimanishi ,@Tapasvi_5fires
For making college life livable
Introducing तत्त्व (𝗧𝗮𝘁𝘃𝗮)
A digital museum dedicated to preserving and connecting ancient Indian texts and traditions, rooted in सनातन धर्म
A step toward the तत्त्व that dwells within
I'm pleased to share that after putting in a lot of efforts and hard work, I have curated the first high quality and clean audio dataset of Shrimad Bhagavad Gita.
I hope this dataset proves to be helpful to all🌸
ॐ नमो भगवते वासुदेवाय 🙏🙏
Sharing the link in the comments 🧵