[TTNN Dialect] Conv2d op missing some optional parameters

[odjuricicTT]

Currently Conv2dOp in TTNN IR is missing a few arguments from ttnn conv2d invoke function.
Mainly Conv2dConfig, DeviceComputeKernelConfig and MemoryConfig (related #620).

TTNN IR Conv2dOp should implement all these arguments to be exactly the same as in the actual op call:

struct Conv2dOperation{
    static Result invoke(
        uint8_t queue_id,
        const ttnn::Tensor& input_tensor,
        const ttnn::Tensor& weight_tensor,
        Device * device,
        uint32_t in_channels,
        uint32_t out_channels,
        uint32_t batch_size,
        uint32_t input_height,
        uint32_t input_width,
        std::array<uint32_t, 2> kernel_size,
        std::array<uint32_t, 2> stride,
        std::array<uint32_t, 2> padding,
        std::array<uint32_t, 2> dilation,
        uint32_t groups,
        std::optional<const ttnn::Tensor> bias_tensor = std::nullopt,
        const std::optional<const Conv2dConfig>& conv_config_ = std::nullopt,
        const std::optional<const DeviceComputeKernelConfig>& compute_config_ = std::nullopt,
        const std::optional<const MemoryConfig>& memory_config = std::nullopt);

https://github.com/tenstorrent/tt-metal/blob/main/ttnn/cpp/ttnn/operations/conv/conv2d/conv2d.hpp#L44

[odjuricicTT]

@sdjordjevicTT Do you think that someone from your team can take a look after the holidays?

[sdjordjevicTT]

Thanks for reporting this @odjuricicTT, @jserbedzijaTT will pick this up after vacation.

[odjuricicTT]

Great, i think that this should be staged in a way so that Conv2dConfig is added first. It's the most important for optimizer usecase.

MemoryConfig is a bit more compilcated and required further discussion.