Metraipolis › archive › dc_fa423da9468d
Giraffe: A Mapping Architecture from Hidden Text Representations to Visual Embeddings for Efficient Graphic Design — arXiv:2608.23970
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.23970
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T06:36:41.078Z
- SHA-256 of the archived text
280b1072c9a1579f91961d0611faba022fcb719a867978e2f69e6853b96f4e04- Stable URL of this page
https://metraipolis.com/documento/dc_fa423da9468d
How to cite
Giraffe: A Mapping Architecture from Hidden Text Representations to Visual Embeddings for Efficient Graphic Design — arXiv:2608.23970 — arXiv (en), https://arxiv.org/abs/2608.23970 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fa423da9468d (Metraipolis archive dc_fa423da9468d, sha256 280b1072c9a1579f91961d0611faba022fcb719a867978e2f69e6853b96f4e04)
BibTeX
@misc{dcfa423da9468d,
title = {Giraffe: A Mapping Architecture from Hidden Text Representations to Visual Embeddings for Efficient Graphic Design — arXiv:2608.23970},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.23970},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fa423da9468d; sha256 280b1072c9a1579f91961d0611faba022fcb719a867978e2f69e6853b96f4e04},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.23970. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Multimodal large language models (MLLMs) have made significant progress in understanding and interpreting mul- timedia content. However, their ability to generate me- dia remains limited. Recent approaches have attempted to bridge this gap by translating the hidden representations of token sequences into the embedding space of visual models or directly into raw image data. However, these methods often represent each image using multiple specialised to- kens which significantly increases the input length. This be- comes a major limitation for tasks such as graphic design generation where the output typically involves a seamless blend of thousands of tokens across text, multiple images, and layout information. To address this challenge, a novel architecture is proposed that maps hidden token represen- tations to the embedding space of visual models, such as CLIP ViT-L/14, using a single [IMG] token per image. The architecture employs two shallow MLP blocks, each with a separate compression module followed by a shared expan- sion module, trained with six distinct loss functions. One block aids the other during training and is omitted during inference, resulting in a lightweight solution. Strong perfor- mance is demonstrated in both image-to-design and text-to- design generation tasks. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-26 in cs.LG). Categorie arXiv: Artificial Intelligence (cs.AI) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2608.23970 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-18T06:28:52.333Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-26?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).