Jia-Bin Huang : vidéos et actus IA

Jia-Bin Huang

Créateursenil y a 13 heures

Vidéo

How Thousands of GPUs Train One AI: 5 Ways They Work Together

Jia-Bin Huang

Créateursenil y a 10 jours

Vidéo

Training Models Too Big for One GPU | ZeRO & FSDP

Jia-Bin Huang

Créateursenil y a 15 jours

Vidéo

Why AI Doesn't Actually Read Words

Jia-Bin Huang

Créateursenil y a 18 jours

Vidéo

What If an AI Could Edit Its Own Memory?

Jia-Bin Huang

Créateursenil y a 23 jours

Vidéo

DeepSeek Splits the Transformer in Two. Here’s Why.

Jia-Bin Huang

Créateursenil y a 2 mois

Vidéo

How LLMs Get Faster Without Changing Their Outputs | Speculative Decoding

Jia-Bin Huang

Créateursenil y a 2 mois

Vidéo

Two API Calls Exposed AI's Hidden Reasoning

Jia-Bin Huang

Créateursenil y a 2 mois

Vidéo

How Kimi Trains 896 Experts Without Collapse

Jia-Bin Huang

Créateursenil y a 3 mois

Vidéo

Jia-Bin Huang

Avis des membres0

Publications

How Thousands of GPUs Train One AI: 5 Ways They Work Together

Training Models Too Big for One GPU | ZeRO & FSDP

Why AI Doesn't Actually Read Words

What If an AI Could Edit Its Own Memory?

DeepSeek Splits the Transformer in Two. Here’s Why.

How LLMs Get Faster Without Changing Their Outputs | Speculative Decoding

Two API Calls Exposed AI's Hidden Reasoning

How Kimi Trains 896 Experts Without Collapse

How Small Models Learn to Think Like Giants | On-Policy Distillation

Jia-Bin Huang

Avis des membres0

Publications

How Thousands of GPUs Train One AI: 5 Ways They Work Together

Training Models Too Big for One GPU | ZeRO & FSDP

Why AI Doesn't Actually Read Words

What If an AI Could Edit Its Own Memory?

DeepSeek Splits the Transformer in Two. Here’s Why.

How LLMs Get Faster Without Changing Their Outputs | Speculative Decoding

Two API Calls Exposed AI's Hidden Reasoning

How Kimi Trains 896 Experts Without Collapse

How Small Models Learn to Think Like Giants | On-Policy Distillation