Sabitlenmiş Tweet
EM
2.6K posts

EM
@ndroyt
Software Engineer + Techie + arT + filmS + yogI + buddhisT + readeR + musiC + vegaN + Español + English + Français + Deutsch
Mexique Katılım Haziran 2009
1.1K Takip Edilen251 Takipçiler

No al Mega Proyecto de ROYAL CARIBBEAN en Cozumel - Detengamos Royal Beach Club - ¡Firma la petición! c.org/bMbYNJD4mY via @Change_Mex
Español

Un científico de Yemen mostró un concepto de prisión de IA del futuro
Cognify propone encerrar a los criminales en cápsulas especiales y reeducar sus cerebros con falsos recuerdos generados por redes neuronales. El resultado son nuevas personas que no querrán violar la ley.
¿Qué dicen ustedes?
Español

@DavidGalanBolsa @condedracco 😂😂😂 que extremistas, solo es un día… eso no define su vida entera
Español

@condedracco Era una de las cosas que llama la atención. Solo sale cocinandole. Lo demás todo ocio individual
Español

@Sophia_Vittar Esta genial que cada quien decida sobre su cuerpo, mujeres, hombres. Si no te atrae algo de una persona pues buscas alguien que se alinea a lo que tu esperas y ya.
Español

Las Ballenas de México están en grave peligro secure.avaaz.org/community_peti…
Español

Trump floats $5 million 'gold card' as a route to U.S. citizenship cnbc.com/2025/02/25/tru…
English
EM retweetledi
EM retweetledi

👩🎨🥐🐀 ¡Hasta nos duele la crepa de la emoción! “Johanne Sacreblu” será proyectada en cines eluniversal.com.mx/tendencias/es-…
Español
EM retweetledi

Cómo lector de Baudelaire y un curso de francés básico, debo de servir que #JohanneSacreblue es una fiel representación de Francia y es arte.
#JohanneSacrebluEnPantallas

Español
EM retweetledi

Porque el público lo pidió y esta obra de AGTE se lo merece 💖 Póster exprés de la heroina nacional, Johanne Sacreblu de @CamilaDAurora 🇫🇷🥖
Gracias a Cam y a todo su equipo por unirnos y recordarnos que el ingenio 100% mexa no tiene igual 🇲🇽✨
#johannesacrebluenpantallas

Español
EM retweetledi

5 text chunking strategies for RAG, visually explained!
In a RAG system, chunking means dividing large documents into smaller, manageable pieces called "chunks."
This process boosts the efficiency and accuracy of retrieval, directly improving the quality of generated responses.
Let's understand them one-by-one! 🚀
(Refer the GIF below as you read...)
1️⃣ Fixed-Size Chunking
This straightforward method splits text into uniform chunks based on a set number of characters, words, or tokens.
For example, dividing a document into chunks of 500 tokens each.
Simple, but it can disrupt the semantic flow!
2️⃣ Semantic Chunking:
Segments text based on meaningful units like sentences, paragraphs, or thematic sections.
Converts each segment into a vector.
Segments with high cosine similarity are combined, and a new chunk is formed when there's a significant shift in context.
3️⃣ Recursive Chunking:
Chunking is first done based on primary separators like paragraphs, or thematic sections.
Each paragraph is then further split into smaller chunks based on a chunk size limit.
It balances chunk size with semantic integrity.
4️⃣ Document Structure-Based Chunking:
Leverages the inherent structure of documents—like headings, sections, or paragraphs—to inform chunk boundaries.
Aligns chunks with the document's logical sections, maintaining structural integrity.
5️⃣ LLM based chunking:
The LLM processes the text and generates meaningful and semantically-isolated statements.
Higher semantic accuracy, as LLM understands context and meaning beyond simple heuristics.
To summarise, here are all 5 explained in a single frame!
In the next tweet, I've share link to a detailed article on this and link to download out FREE Data Science PDF covering 150+ essentials lessons in DS/ML.
_____
Find me → @akshay_pachaar✔️
For more insights and tutorials on AI and Machine Learning!
GIF
English
EM retweetledi

Choosing the right machine learning model!
#AI #MachineLearning #DeepLearning #DataScience #NLP #NeuralNetworks #GenerativeAI #LLM #Python #Code #100DaysOfCode @DataScienceDojo
@SpirosMargaris @PawlowskiMario @mvollmer1 @gvalan @ipfconline1 @LaurentAlaus @Shi4Tech @Fisher85M @kalydeoo @Ym78200 @Nicochan33 @Fabriziobustama @3itcom @chidambara09 @Analytics_699 @Khulood_Almani @tewoz @ahier @EvanKirstel @rwang0 @sallyeaves @helene_wpli @FrRonconi @DigitalColmer @arielSTRABONI @HaroldSinnott @fogle_shane @rshevlin @jeffkagan @jeancayeux @RLDI_Lamy @CurieuxExplorer

English
EM retweetledi

Microsoft just changed the game! 🔥
They've open-sourced bitnet.cpp: a blazing-fast 1-bit LLM inference framework that runs directly on CPUs.
Why is this a game-changer❓
You can now run 100B parameter models on local devices with up to 6x speed improvements and 82% less energy consumption—all without a GPU!
The future we've been waiting for: fast, efficient, and private AI that works anytime, anywhere.✨
Link to the GitHub repo in next tweet!
_____
Find me → @akshay_pachaar ✔️
For more insights and tutorials on AI and Machine Learning!
English
EM retweetledi

¡Que Nestlé pague precios justos a Cafeticultores de Chiapas, México! - ¡Firma la petición! chng.it/RFZMTngdNp via @Change_Mex
Español








