Expired50% confidenceFactTime unknown
In SD1.5, the CLIP text encoder converts prompts into 768-dimensional vectors, while in SDXL they are converted into 1024-dimensional vectors.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026(expired)
Sources
Stable Diffusion Local Deployment Guide: Run AI Image Generation Free with 8GB RAM
bilibiliAI画梦师-由乃
Related Claims
Unverified文本条件注入通过CLIP的交叉注意力层实现，ViT-L/14版本的CLIP将文本token映射为768维向量序列73% similarUnverifiedCLIP's text encoder is based on the Transformer architecture and maps input text to 512-dimensional or 768-dimensional embedding vectors.72% similarUnverified在Stable Diffusion中，CLIP文本编码器将自然语言切分为最多77个token，每个token映射为768维或1024维嵌入向量67% similarVerifiedtext-embedding-ada-002将文本转化为1536维的稠密向量，语义相近的文本在向量空间中距离更近65% similarVerifiedStable Diffusion's text guidance converts prompts into vectors through the CLIP text encoder.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/59732API
curl https://kongchang.com/api/v1/knowledge/claims/59732MCP
get_claim(id=59732)