yamatazen/Ayla-Light-Extra
Same weights, repacked vs yamatazen/Ayla-Light-12B-Stock. Same weights in different files (re-saved, re-sharded or new metadata), plus 72 new tensors.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/81, MLP 0/120, Attention 0/160
2 derivatives checked
- mradermacher/Ayla-Light-Extra-i1-GGUF: Quantization
- mradermacher/Ayla-Light-Extra-GGUF: Quantization