, *optional*, defaults to `False`):
            Whether to quantize the model or not.
        force_dequant (`str`, *optional*, defaults to `"none"`):
            Force dequantize specific nonlinear layer. Dequatized layers are then executed with full precision.
            `"none"`, `"gelu"`, `"softmax"`, `"layernorm"` and `"nonlinear"` are supported. As deafult, it is set as
            `"none"`, which does not dequantize any layers. Please specify `"gelu"`, `"softmax"`, or `"layernorm"` to
            dequantize GELU, Softmax, or LayerNorm, respectively. `"nonlinear"` will dequantize all nonlinear layers,
            i.e., GELU, Softmax, and LayerNorm.
    Z