Rombo-Org/reka-flash-3-GGUF_QX_k_Bf16

Quantization vs RekaAI/reka-flash-3. Rombo-Org/reka-flash-3-GGUF_QX_k_Bf16 is a GGUF quantization (Q3_K_BF16) of the same weights: 91 tensors match exactly and 308 within quantization error.

Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/89, MLP 0/132, Attention 0/176

On the Hub · Downloads on Model Pulse