roskosmos19/Orca-4B-Thinking
Different architecture vs Qwen/Qwen3.5-4B. Different architecture: only 89 of 738/398 tensors match by name and shape.
2 derivatives checked
- mradermacher/Whale-4B-Thinking-i1-GGUF: Quantization
- mradermacher/Whale-4B-Thinking-GGUF: Same weights, converted
10 other repos hold byte-for-byte identical weight files.