DavidAU/Reka-Flash-3-21B-Reasoning-Uncensored-MAX-NEO-Imatrix-GGUF
Quantization vs RekaAI/reka-flash-3. DavidAU/Reka-Flash-3-21B-Reasoning-Uncensored-MAX-NEO-Imatrix-GGUF is a GGUF quantization (Q8_0/BF16) of the same weights: 90 tensors match exactly and 309 within quantization error.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/89, MLP 0/132, Attention 0/176