doth4580/Qwen3.8-27B-DFlash2-NVFP4-RTNcal-FIXED
Same weights, repacked (fewer layers) vs incoai/Qwen3.8-27B-DFlash2. doth4580/Qwen3.8-27B-DFlash2-NVFP4-RTNcal-FIXED matches part of incoai/Qwen3.8-27B-DFlash2 (46 of its 81 tensors; the rest of incoai/Qwen3.8-27B-DFlash2 has no counterpart). On the shared part: Same weights in different files (re-saved, re-sharded or new metadata).
Tensors changed: other 0/4, Norms 0/22, Attention 0/15, MLP 0/5