z-lab/Qwen3.5-27B-DFlash
Different architecture vs Qwen/Qwen3.5-27B. Different architecture: only 33 of 1199/69 tensors match by name and shape.
2 derivatives checked
- Anbeeld/Qwen3.5-27B-DFlash-GGUF: Same weights, converted
- AtomicChat/Qwen3.5-27B-DFlash-GGUF: Quantization
1 other repos hold byte-for-byte identical weight files.