Feed Entities Podcasts Deep Dives Developers About

[KongchangAI]

Real-time AI knowledge base. Every claim is traceable to its source, with data freshness under 24 hours.

Powered by 240+ first-hand AI sources

Knowledge

Feed
Hot Entities
Podcasts

Developers

CLI
MCP Server
REST API
About

© 2026 KongchangAI. The freshest, most accurate AI knowledge

Unverified50% confidenceBenchmarkExact time

实测在RTX 3060（6GB显存）、32GB RAM上运行预量化Krea2 Turbo，1024×1024分辨率8步Euler采样，速度为1.78 s/it，总耗时17.64秒

1

Sources

50%

Confidence

Medium-term (~90 days)

Relevance

7/12/2026

First Seen

Valid until: 10/10/2026

Sources

ComfyUI-INT4-Fast：6GB显卡流畅运行Flux大模型实测

redditr/StableDiffusion7/10/2026

Related Claims

Unverified开发者公布在RTX 3050 Laptop（4GB）上微调Llama-3.1-8B的实测结果为训练速度119.6 tok/s，峰值显存3.32 GB80% similar UnverifiedRTX 5090上Qwen3.6 27B模型预填充速度从174提升到253 token/s，显存占用约18GB76% similar Unverified6GB显存的RTX 3060可以流畅运行基于Flux的Krea2 Turbo大模型74% similar UnverifiedRTX 5090上Qwen3.6 35B-A3B推理速度达220 token/s，显存占用约23GB73% similar UnverifiedLlama 3.1 8B Instruct在配备RTX 3090（24GB显存）的服务器上即可流畅运行，推理延迟通常在数秒以内73% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/489327

API

curl https://kongchang.com/api/v1/knowledge/claims/489327

MCP

get_claim(id=489327)