…for NPU

…hon deps

This reverts commit 67389d6.

…openvino.genai into guozhong/whisper_pipeline

…or tensors initialization, compression, alpha-only mode, better diagnostics about errors.

Reapplying openvinotoolkit#1161 to 2024.5 release branch

…component sample

…oolkit#1169)

This reverts commit e473381.

…plementation of unet for initial NPU support (openvinotoolkit#1101)

For now, NPU can be used like this. The unet model must be reshaped to a
static shape before compile is invoked.

```cpp
std::filesystem::path root_dir = models_path;
auto pipe = ov::genai::Text2ImagePipeline::stable_diffusion(
    ov::genai::Text2ImagePipeline::Scheduler::from_config(root_dir / "scheduler/scheduler_config.json"),
    ov::genai::CLIPTextModel(root_dir / "text_encoder", "CPU"),
    ov::genai::UNet2DConditionModel(root_dir / "unet")
        .reshape(2, 512, 512, 77)
        .compile("NPU", ov::cache_dir("./cache")),
    ov::genai::AutoencoderKL(root_dir / "vae_decoder", "GPU", ov::cache_dir("./cache")));
ov::Tensor image = pipe.generate(prompt);
```

…1162)

Fixing the regression introduced in
openvinotoolkit#1118
CVS-156870

---------

Co-authored-by: Chen Peter <peter.chen@intel.com>

Fix the issue when running stable-diffusion

[ INFO ] Traceback (most recent call last):
File
"/home/guozhong/openvino.genai_2024_5/openvino.genai/tools/llm_bench/benchmark.py",
line 202, in main
iter_data_list, pretrain_time, iter_timestamp =
CASE_TO_BENCH[model_args['use_case']](
File
"/home/guozhong/openvino.genai_2024_5/openvino.genai/tools/llm_bench/task/image_generation.py",
line 198, in run_image_generation_benchmark
image_gen_fn(image_param, num, prompt_idx_list[image_id], pipe, args,
iter_data_list, proc_id, mem_consumption)
File
"/home/guozhong/openvino.genai_2024_5/openvino.genai/tools/llm_bench/task/image_generation.py",
line 50, in run_image_generation
f'steps={input_args["num_inference_steps"]},
width={input_args["width"]}, height={input_args["height"]},
guidance_scale={input_args["guidance_scale"]}'
KeyError: 'guidance_scale'

---------

Co-authored-by: Ilya Lavrenov <ilya.lavrenov@intel.com>