vidore/colqwen2.5omni-base

Same weights, repacked vs Qwen/Qwen2.5-Omni-3B. Same weights in different files (re-saved, re-sharded or new metadata), plus 2 new tensors.

Tensors changed: other 0/7, MLP 0/381, Norms 0/233, Attention 0/613, Embeddings 0/1

1 derivatives checked