autotrust/JEV-27B

Same weights, repacked vs Qwen/Qwen3.8-27B. Same weights in different files (re-saved, re-sharded or new metadata).

Tensors changed: Norms 0/209, Attention 0/448, MLP 0/192, Embeddings 0/1, lm_head 0/1

2 derivatives checked

• prithivMLmods/JEV-27B-GGUF: Same weights, converted
• webmp3/Sakura-Humanlike-Chat-2.0-JEV-AutoRound-GGUF: Different architecture

On the Hub · Downloads on Model Pulse