weitusi/lianren-4b
Same weights, repacked vs Qwen/Qwen3.5-4B. Same weights in different files (re-saved, re-sharded or new metadata), plus 32 new tensors.
Tensors changed: Embeddings 0/5, MLP 0/199, Attention 0/324, Norms 0/209, other 0/1
Same weights, repacked vs Qwen/Qwen3.5-4B. Same weights in different files (re-saved, re-sharded or new metadata), plus 32 new tensors.
Tensors changed: Embeddings 0/5, MLP 0/199, Attention 0/324, Norms 0/209, other 0/1