How do computers know which words matter most in a text? 🤔
TF-IDF is a classic NLP metric that still works perfectly. I just published a simple guide covering the core idea, the math, and a step-by-step example.
📖 Check it out: https://t.co/0aHCIthUxp
#NLP#TextProcessing
Still impressed by Cassandra's performance: 3.7B writes at 0.014ms, 51M reads at 0.30ms on my 12-node AWS cluster. Rack-aware across 3 AZs. Proper modeling makes all the difference for distributed systems.
#Cassandra#NoSQL#DistributedSystems#AWS
When the conference date is approaching and you need to finish all the fine-tuning experiments asap. There’s no time to sleep. 😴
#ai#finetuning#deeplearning#llm
@mazei_lucas Eu estou testando o n8n para colocar em prod uma RAG. Realmente facilita muito a vida, a facilidade de conectar em diversas bases e escolher entre vários modelos.
Ainda gosto de criar na mão por conta de pesquisas tbm, mas hoje para o ambiente corporativo é essencial.
Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning
They add a search engine into the DeepSeek-R1 GRPO based RL training recipe and find that it achieves SOTA.