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: 98be1ca02abea0daf9b002f563c828683310f255c4094c0d2f0c0c35c7f05778