specific method · filed under model architecture
Joint multimodal decoding in a shared hidden space
Modalities are projected into a shared hidden space and processed jointly by the decoder.
- sources
- 3
- model
- 1
- lab adopt it
- 1
- strongest
- core
How sources treat it
One count per evidence span, weakest treatment to strongest.
core 3
Documented in
Further reading
Picked by hand, not extracted: where to read more, not evidence for anything on this page.
- Chameleon: Mixed-Modal Early-Fusion Foundation Models paper arxiv.orgmodalities projected into one shared token/hidden space
Evidence
3 spans quoted from the sources, strongest treatment first.
with all modalities projected into a shared hidden space and processed jointly by the decoder
coremodel architecturein InklingThinking Machines Lab
with all modalities projected into a shared hidden space and processed jointly by the decoder
coremodel architecturein InklingThinking Machines Lab
with all modalities projected into a shared hidden space and processed jointly by the decoder
coremodel architecturein InklingThinking Machines Lab
Filed alongside
Other methods under model architecture :: multimodal architecture.
Encoder-free multimodal architectureEarly fusion multimodal trainingDiscrete token encoding for audioHierarchical patch encoderMixed-modality trainingConfigurable visual token budgetNative multimodal understanding3×3 pixel-unshuffle downsamplingJoint vision-language pre-trainingRaw audio projection into the LLM embedding spaceVision Transformer (ViT)2×2 pixel-shuffle downsamplingAudio encoder initialized from MiMo-AudioAudio Transformer (AuT)Causal streaming ConvNet codec decoderData-parallel-first multimodal encodingDecoupled Encoder Process (DEP)Dedicated multimodal encodersDisaggregated encoder trainingDiscarding the LLM after vision-encoder trainingDual-format coordinate supervisionExplicit text-string timestampsFour-frame audio patches with within-patch bidirectional self-attentionFrom-scratch vision-encoder training with next-token prediction