implementation detail · filed under model architecture
2D coordinate-based positional embeddings
Positional embeddings based on 2D coordinates that are added directly to patch representations.
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
Spatial awareness is maintained by adding 2D coordinate-based positional embeddings directly to the patch representations before a final LayerNorm layer
usedmodel architecturein Gemma 4 12BGoogle DeepMind
Filed alongside
Other methods under model architecture :: positional encoding.
YaRNRotary Position Embedding2D rotary position embeddingRoPE scalingLearned input-dependent relative position biasNo Position EncodingProportional Rotary Position EmbeddingRelative attentionGated attention with partial RoPEOmitting RoPE in attention layersPartial RoPEPer-layer-type rotary position scalesTemporal-Modality Rotary Position Embedding