0 rather than building a new multimodal architecture.
Apertus adds image, audio and longer reasoning without architectural changes
The authors extend two fully open language models through continued multimodal training, while preserving their architecture and publishing the resulting weights, data pipelines and recipes.
Independent
Project Apertus (ETH Zurich) · : (ETH Zurich) · Yixuan Xu (ETH Zurich) · Yassine Turki (EPFL) · Vinko Sabolčec (EPFL) · Valentina Pyatkin (ETH Zurich) · +127 more
Research Digest··3 min read
Project Apertus upgrades its 8-billion and 70-billion parameter text models to process images and audio, use contexts of up to 262,144 tokens and optionally produce explicit reasoning traces.
Why this paper
Independent
In one line
Apertus 1.5 extends open text models to multimodal reasoning, adding image and audio understanding with thinking mode.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors
- ✓Reports numbers on named benchmarks (3 benchmarks)
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§