By expanding the token length, Long-CLIP can process longer text inputs more effectively, capturing more context and details. This is particularly useful for generating images from detailed descriptions, as it allows the model to consider a broader range of information, resulting in higher-quality outputs.

Longclip-L

CLIP-GmP-ViT-L-14 Text - for improved detailed in images containing text.

Files with SHA256: 0830f1119276e3f3e6ee250dd1bf5e4d84b61ceb8d7d1d56c83601efc2a00950