FIT17/Confucius4-R2T2-GGUF
Same weights, converted (fewer layers) vs netease-youdao/Confucius4-R2T2. FIT17/Confucius4-R2T2-GGUF matches part of netease-youdao/Confucius4-R2T2 (310 of its 707 tensors; the rest of netease-youdao/Confucius4-R2T2 has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → BF16/F32).
Tensors changed: Embeddings 0/1, Norms 0/113, MLP 0/84, Attention 0/112