quantization for own model

Can I perform the quantization using the "optimum" for my own executorch FP16 model, which needs to be quantized as q4fp16? Here, my model includes mamba blocks. 
If yes, then can you guide with step by step process?