AISatoshi (@AiXsatoshi)Mistral의 Mistral-Medium-3.5-128B-EAGLE에 대해 speculative decoding 가속 테스트 결과를 공유했다. 비디오 코딩 기준 수용률이 약 25~30% 수준이며, MoE 모델이 빠르지만 Dense 모델과 전용 speculative decoding 모델 조합도 유용하다고 언급했다.https://x.com/AiXsatoshi/status/2049543302530355622#mistral #speculativedecoding #llm #aigeneration #moe
Related
'replace workers, lower costs, increase profits, and avoid being left behind by competitors. But when every firm does th...
'replace workers, lower costs, increase profits, and avoid being left behind by competitors. But when every firm does the same thing, they begin to undercut the economy on which th...
AI · Anthropic beats OpenAI in revenue for the first timeAnthropic's Q2 sales hit $11.6 billion, passing OpenAI's $6.7 b...
AI · Anthropic beats OpenAI in revenue for the first timeAnthropic's Q2 sales hit $11.6 billion, passing OpenAI's $6.7 billion, as both race toward IPOs.Read the rest at: https://t...
🌪️ DeepMind acaba de publicar en Nature un modelo de IA que da un día extra de aviso en predicción de ciclones.WeatherNe...
🌪️ DeepMind acaba de publicar en Nature un modelo de IA que da un día extra de aviso en predicción de ciclones.WeatherNext Cyclones:• Trayectoria + intensidad + tamaño del viento• ...