add channel wise quantization option for QDQ, and opt for intel NPU

[bopeng1234]

Description

add 4bits channel-wised quantization capability for DequantizeLinear Op for phi3 model, it optimized the TPS on Intel NPU

JIRA - https://jira.devtools.intel.com/browse/EISW-163602

Motivation and Context

As Intel's NPU support to LLM shows https://github.com/openvinotoolkit/openvino.genai/tree/master/samples/python/text_generation#npu-support

if we want to use onnx quantized model in intel NPU, like phi3, the quantized model need to meet two requirements:

1. symmetric, zp=0
2. channel wised quantize, block_size = K

So this PR's changes is to enable the channel wised quantize, and symmetric.

we tested it with onnxruntime_genai changes (we created a PR to onnxruntime_genai too to support this extra args, microsoft/onnxruntime-genai#1362). and openvino changes openvinotoolkit/openvino#30265

command:
python -m onnxruntime_genai.models.builder -o E:\download\onnx\Phi-3-mini-4k-instruct-onnx-channelwise-modified-QDQ-T -p int4 -e cpu -i E:\download\huggingface\Phi-3-mini-4k-instruct --extra_options use_channel_wised_quantization=1 use_qdq=1

normally without channel-wised quantize model, the phi3 with NPUW, runs about 4000ms per token when kv cache model.
apply this PR, and phi3 with NPUW, runs about 150ms per token, speed up 20+ times.

[bopeng1234]

@ankitm3k , create this new one, only add the QDQ CW, removed Qoperater related code.

[ankitm3k]

@bopeng1234 kindly resolve conflicts

[ankitm3k]

Reviewed & tested the changes, LGTM