PTQ signature is too big

Current PTQ signature is a little overwhelming:

void quantizeNetwork(std::shared_ptr<GraphView> graphView, 
        std::uint8_t nbBits,
        std::vector<std::shared_ptr<Tensor>> inputDataSet,
        Clipping clippingMode,
        DataType targetType,
        bool noQuant,
        bool optimizeSigns,
        bool singleShift,
        bool useCuda,
        bool foldGraph,
        bool bitshiftRounding,
        bool verbose);

Proposed modifications:

• verbose: Should be deprecated in favor of aidge log system
• useCuda: Should be removed, using cuda should be decided locally
• foldGraph: Shouldn't be a parameter of attribute use should run it afterward but the PTQ method should not handle it
• nbBits: isn't it redundant with targetType?
• noQuant: this parameter seems badly named, what is it even used for?
• optimizeSigns: More documentation is required I don't understand what is does ...
• singleShift: How are we gonna do when we introduce fixed point? Add another boolean? I think we should create an enum ScalingType and use it as an argument instead
• bitshiftRounding: I would be in favor to remove this argument and use the proposed ScalingType enum instead.