metadata
base_model: RekaAI/reka-flash-3
EXL3 quantization of reka-flash-3, 3 bits per weight.
HumanEval (argmax)
Model | Q4 | Q8 | FP16 |
---|---|---|---|
reka-flash-3-exl3-3bpw | 87.8 | 90.2 | 90.9 |
reka-flash-3-exl3-4bpw | 89.0 | 88.4 | 87.2 |
base_model: RekaAI/reka-flash-3
EXL3 quantization of reka-flash-3, 3 bits per weight.
Model | Q4 | Q8 | FP16 |
---|---|---|---|
reka-flash-3-exl3-3bpw | 87.8 | 90.2 | 90.9 |
reka-flash-3-exl3-4bpw | 89.0 | 88.4 | 87.2 |