### Install Sample Binaries

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/tools/continuous_batching/accuracy/CMakeLists.txt

Installs the defined targets to the 'samples_bin/' directory as runtime components. This makes the compiled executables accessible after installation.

```cmake
install(TARGETS ${TARGET_NAME} ${TARGET_NAME_CB}
        RUNTIME DESTINATION samples_bin/
        COMPONENT tools_bin
        EXCLUDE_FROM_ALL)
```

--------------------------------

### Install Dependencies

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/site/README.md

Run this command to install all project dependencies.

```bash
$ npm i
```

--------------------------------

### Install Dependencies for Running Samples

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/python/video_generation/README.md

Installs necessary packages to run the OpenVINO GenAI samples. Use the --upgrade-strategy eager option to ensure the latest versions are installed.

```sh
pip install --upgrade-strategy eager -r ../../deployment-requirements.txt
```

--------------------------------

### Example Text to Video Generation

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/video_generation/README.md

Provides a concrete example of running the text-to-video sample with the LTX-Video model and a specific prompt.

```bash
./text2video ltx_video_ov/FP32 "A woman with long brown hair and light skin smiles at another woman with long blonde hair"
```

--------------------------------

### Install Executable

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/whisper_speech_recognition/CMakeLists.txt

Installs the 'whisper_speech_recognition' executable to the 'samples_bin/' directory at runtime. This ensures the sample application is placed in the correct location after installation.

```cmake
install(TARGETS whisper_speech_recognition
        RUNTIME DESTINATION samples_bin/
        COMPONENT samples_bin
        EXCLUDE_FROM_ALL)
```

--------------------------------

### Install WWB from Source

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/tools/who_what_benchmark/README.md

Install the WWB tool from its source directory using pip. This command installs the core package and its required dependencies.

```bash
pip install .
```

--------------------------------

### Install Node.js dependencies

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/js/speech_generation/README.md

Install project dependencies from the samples/js directory.

```bash
npm install
```

--------------------------------

### Start Local Development Server

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/site/README.md

Starts a local development server that reflects changes live without requiring a restart.

```bash
$ npm run start
```

--------------------------------

### Set Up Environment with setupvars.sh

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/src/docs/BUILD.md

Source the OpenVINO setupvars.sh script to configure the build environment. Replace <INSTALL_DIR> with your OpenVINO installation path.

```sh
source <INSTALL_DIR>/setupvars.sh
```

--------------------------------

### Run Text to Image Generation

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/image_generation/README.md

Executes the text-to-image generation sample with a specified prompt and model path. Refer to 'Get Started with Samples' for detailed instructions.

```sh
stable_diffusion ./dreamlike_anime_1_0_ov/FP16 'cyberpunk cityscape like Tokyo New York with tall buildings at dusk golden hour cinematic lighting'
```

--------------------------------

### Install OpenVINO GenAI

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/README.md

Install the OpenVINO GenAI library using pip. This is the first step to using the library.

```sh
pip install openvino-genai
```

--------------------------------

### Install OpenVINO GenAI Artifacts

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/src/docs/BUILD.md

Install the built artifacts to the specified installation directory.

```sh
cmake --install ./build/ --config Release --prefix <INSTALL_DIR>
```

--------------------------------

### Install Hugging Face Hub

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/site/docs/guides/model-preparation/download-openvino-models.mdx

Install the huggingface_hub package to enable model downloads from Hugging Face.

```bash
pip install huggingface_hub
```

--------------------------------

### Install WWB with Nightly Builds

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/tools/who_what_benchmark/README.md

Install WWB along with nightly builds of OpenVINO, openvino-tokenizers, and openvino-genai. This is useful for testing the latest features and fixes.

```bash
PIP_PRE=1 \
PIP_EXTRA_INDEX_URL=https://storage.openvinotoolkit.org/simple/wheels/nightly \
pip install .
```

--------------------------------

### Install ModelScope and Optimum Intel

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/site/docs/guides/model-preparation/convert-to-openvino.mdx

Install the `modelscope` and `optimum-intel` packages to facilitate downloading models from ModelScope and converting them to OpenVINO IR format.

```bash
pip install modelscope --requirement ./samples/export-requirements.txt
```

--------------------------------

### Basic Text Generation in C++

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/src/README.md

A minimal example demonstrating how to initialize an LLMPipeline and generate text using a simple prompt.

```cpp
#include "openvino/genai/llm_pipeline.hpp"
#include <iostream>

int main(int argc, char* argv[]) {
    std::string models_path = argv[1];
    ov::genai::LLMPipeline pipe(models_path, "CPU");
    std::cout << pipe.generate("The Sun is yellow because", ov::genai::max_new_tokens(256));
}
```

--------------------------------

### Install OpenVINO GenAI and Dependencies

Source: https://context7.com/openvinotoolkit/openvino.genai/llms.txt

Installs the OpenVINO GenAI library and necessary tools for model conversion and downloading pre-converted models.

```bash
pip install openvino-genai
```

```bash
pip install optimum[openvino]
optimum-cli export openvino --model TinyLlama/TinyLlama-1.1B-Chat-v1.0 --weight-format int4 --trust-remote-code TinyLlama_1_1b_v1_ov
```

```bash
pip install huggingface-hub
huggingface-cli download OpenVINO/TinyLlama-1.1B-Chat-v1.0-int4-ov --local-dir TinyLlama_1_1b_v1_ov
```

--------------------------------

### Prepare Python Virtual Environment

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/tools/llm_bench/README.md

Commands to create a virtual environment, install dependencies, and clone the repository.

```bash
python3 -m venv ov-llm-bench-env
source ov-llm-bench-env/bin/activate
pip install --upgrade pip

git clone  https://github.com/openvinotoolkit/openvino.genai.git
cd openvino.genai/tools/llm_bench
pip install -r requirements.txt
```

```bash
pip install -U --upgrade-strategy eager -r requirements.txt
```

--------------------------------

### Install timm and einops for InternVL2

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/site/docs/supported-models/index.mdx

Required packages for converting InternVL2 models. Ensure these are installed before proceeding with the conversion.

```bash
pip install timm einops
```

--------------------------------

### Install OpenVINO GenAI via npm

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/site/docs/getting-started/installation.mdx

Use this command to install the Node.js bindings for OpenVINO GenAI.

```bash
npm install openvino-genai-node
```

--------------------------------

### Set up Python Virtual Environment

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/tools/who_what_benchmark/README.md

Create and activate a Python virtual environment for installing WWB and its dependencies. This is a recommended step for managing project dependencies.

```bash
python -m venv eval_env
source eval_env/bin/activate
```

--------------------------------

### Install WWB with Extra Requirements

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/tools/who_what_benchmark/README.md

Install WWB with specific extra requirements, such as support for certain model configurations like 'minicpm-o-2_6'. Refer to setup.py for available extras.

```bash
pip install .[minicpm-o-2_6]
```

--------------------------------

### Install Dependencies and Export Model to OpenVINO IR

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/c/text_generation/README.md

Install necessary dependencies using pip and then export a model to OpenVINO IR format using the optimum-cli. The --upgrade-strategy eager option ensures optimum-intel is updated.

```sh
pip install --upgrade-strategy eager -r ../../export-requirements.txt
optimum-cli export openvino --model <model> <output_folder>
```

--------------------------------

### Install macOS command line tools

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/src/docs/BUILD.md

Install the necessary Clang compiler and command line tools via Xcode.

```sh
xcode-select --install
```

--------------------------------

### Example LoRA Text to Video Generation

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/video_generation/README.md

Demonstrates running the LoRA text-to-video sample with the LTX-Video model, a prompt, and a specific LoRA adapter with an alpha value.

```bash
./lora_text2video ltx_video_ov/FP32 "In the style of Pixel, the video shifts to a majestic castle under a starry sky." ltx_pixel_pytorch_lora_weights.safetensors 3.0
```

--------------------------------

### Install Dependencies and Export Model

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/image_generation/README.md

Installs necessary dependencies and exports a Stable Diffusion model to OpenVINO format using the optimum-cli. Ensure optimum-intel is up-to-date.

```sh
pip install --upgrade-strategy eager -r ../../requirements.txt
optimum-cli export openvino --model dreamlike-art/dreamlike-anime-1.0 --task stable-diffusion --weight-format fp16 dreamlike_anime_1_0_ov/FP16
```

--------------------------------

### Install Dependencies and Export Model

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/js/whisper_speech_recognition/README.md

Install necessary Python packages and export the Whisper model to OpenVINO format using the optimum-cli.

```sh
pip install --upgrade-strategy eager -r <GENAI_ROOT_DIR>/samples/requirements.txt
optimum-cli export openvino --trust-remote-code --model openai/whisper-base whisper-base
```

--------------------------------

### Define Sample Executables

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/visual_language_chat/CMakeLists.txt

Defines build targets for various VLM and chat samples, linking them against OpenVINO GenAI and setting installation properties.

```cmake
add_executable(visual_language_chat visual_language_chat.cpp load_image.cpp)
target_include_directories(visual_language_chat PRIVATE "${CMAKE_BINARY_DIR}")
target_link_libraries(visual_language_chat PRIVATE openvino::genai)

set_target_properties(visual_language_chat PROPERTIES
    # Ensure out of box LC_RPATH on macOS with SIP
    INSTALL_RPATH_USE_LINK_PATH ON)

install(TARGETS visual_language_chat
        RUNTIME DESTINATION samples_bin/
        COMPONENT samples_bin
        EXCLUDE_FROM_ALL)
```

```cmake
add_executable(visual_language_lora visual_language_lora.cpp load_image.cpp)
target_include_directories(visual_language_lora PRIVATE "${CMAKE_BINARY_DIR}")
target_link_libraries(visual_language_lora PRIVATE openvino::genai)

set_target_properties(visual_language_lora PROPERTIES
    # Ensure out of box LC_RPATH on macOS with SIP
    INSTALL_RPATH_USE_LINK_PATH ON)

install(TARGETS visual_language_lora
        RUNTIME DESTINATION samples_bin/
        COMPONENT samples_bin
        EXCLUDE_FROM_ALL)
```

```cmake
add_executable(encrypted_model_vlm encrypted_model_vlm.cpp load_image.cpp)
target_include_directories(encrypted_model_vlm PRIVATE "${CMAKE_BINARY_DIR}")
target_link_libraries(encrypted_model_vlm PRIVATE openvino::genai)

set_target_properties(encrypted_model_vlm PROPERTIES
    # Ensure out of box LC_RPATH on macOS with SIP
    INSTALL_RPATH_USE_LINK_PATH ON)

install(TARGETS encrypted_model_vlm
        RUNTIME DESTINATION samples_bin/
        COMPONENT samples_bin
        EXCLUDE_FROM_ALL)
```

```cmake
add_executable(benchmark_vlm benchmark_vlm.cpp load_image.cpp ../text_generation/read_prompt_from_file.cpp)
target_include_directories(benchmark_vlm PRIVATE "${CMAKE_BINARY_DIR}")
target_link_libraries(benchmark_vlm PRIVATE openvino::genai cxxopts::cxxopts)
set_target_properties(benchmark_vlm PROPERTIES
    # Ensure out of box LC_RPATH on macOS with SIP
    INSTALL_RPATH_USE_LINK_PATH ON)

install(TARGETS benchmark_vlm
        RUNTIME DESTINATION samples_bin/
        COMPONENT samples_bin
        EXCLUDE_FROM_ALL)
```

```cmake
add_executable(video_to_text_chat video_to_text_chat.cpp)
target_include_directories(video_to_text_chat PRIVATE "${CMAKE_BINARY_DIR}")
ov_genai_link_opencv(video_to_text_chat core imgproc videoio imgcodecs)
target_link_libraries(video_to_text_chat PRIVATE openvino::genai cxxopts::cxxopts)

install(TARGETS video_to_text_chat
        RUNTIME DESTINATION samples_bin/
        COMPONENT samples_bin
        EXCLUDE_FROM_ALL)
```

--------------------------------

### Install Whisper Sample Executable

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/c/whisper_speech_recognition/CMakeLists.txt

Installs the compiled 'whisper_speech_recognition_c' executable into the 'samples_bin/' directory within the runtime destination. It is marked as a component 'samples_bin' and excluded from all targets.

```cmake
install(TARGETS whisper_speech_recognition_c
        RUNTIME DESTINATION samples_bin/
        COMPONENT samples_bin
        EXCLUDE_FROM_ALL)
```

--------------------------------

### Install Optimum Intel for Model Conversion

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/site/docs/guides/model-preparation/convert-to-openvino.mdx

Install the `optimum-intel` package along with its requirements to enable model conversion and optimization. Some models may require additional dependencies listed in `export-requirements.txt`.

```bash
pip install --requirement ./samples/export-requirements.txt
```

--------------------------------

### Initialize and use TextRerankPipeline

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/js/rag/README.md

Example of initializing the rerank pipeline and reranking documents based on a query.

```js
import { TextRerankPipeline } from 'openvino-genai-node';

const pipeline = await TextRerankPipeline(modelPath, { device: "CPU" });

const rerankResult = await pipeline.rerank(query, documents);
```

--------------------------------

### Run Structured Output Generation Sample

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/text_generation/README.md

Execute this command to start the interactive dialog for structured output generation. You can then prompt the model to solve equations step by step.

```bash
structured_output_generation <MODEL_DIR>
```

--------------------------------

### Get Timestamps for Transcription

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/python/whisper_speech_recognition/README.md

Enable timestamp prediction during transcription by setting `return_timestamps=True`. The result object will contain chunks with start and end timestamps.

```python
raw_speech = read_wav("how_are_you_doing_today.wav")
result = pipe.generate(raw_speech, return_timestamps=True)

for chunk in result.chunks:
    print(f"timestamps: [{chunk.start_ts:.2f}, {chunk.end_ts:.2f}] text: {chunk.text}")
# timestamps: [0.00, 2.00] text:  How are you doing today?
```

--------------------------------

### Build OpenVINO GenAI Samples for Windows (Command Prompt)

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/c/text_generation/README.md

Set up the environment and build the samples for Windows using the Command Prompt. Ensure you have sourced the setupvars batch script.

```bat
<INSTALL_DIR>\setupvars.bat
<INSTALL_DIR>\samples\c\build_samples_msvc.bat
```

--------------------------------

### Image2ImagePipeline Generation Example

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/site/docs/use-cases/image-generation/_sections/_run_model/_image2image_python.mdx

Use this snippet to generate an image based on a text prompt and an input image using the Image2ImagePipeline. Ensure you have the necessary libraries (openvino_genai, openvino, PIL, numpy) installed and model_path and prompt variables defined.

```python
import openvino_genai as ov_genai
import openvino as ov
from PIL import Image
import numpy as np

def read_image(path: str) -> ov.Tensor:
    pic = Image.open(path).convert("RGB")
    image_data = np.array(pic)[None]
    return ov.Tensor(image_data)

input_image_data = read_image("input_image.jpg")

pipe = ov_genai.Image2ImagePipeline(model_path, "${props.device || 'CPU'}")
image_tensor = pipe.generate(prompt, image=input_image_data, strength=0.8)

image = Image.fromarray(image_tensor.data[0])
image.save("image.bmp")

```

--------------------------------

### Build OpenVINO GenAI C Samples

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/c/whisper_speech_recognition/README.md

Commands to initialize the environment and build the C samples across different operating systems.

```sh
source <INSTALL_DIR>/setupvars.sh
./<INSTALL_DIR>/samples/c/build_samples.sh
```

```cmd
<INSTALL_DIR>\setupvars.bat
<INSTALL_DIR>\samples\c\build_samples_msvc.bat
```

```powershell
.<INSTALL_DIR>\setupvars.ps1
<INSTALL_DIR>\samples\c\build_samples.ps1
```

--------------------------------

### Verify OpenVINO GenAI Installation

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/src/docs/BUILD.md

Verify the installation of OpenVINO GenAI by importing the library and printing its version. This confirms that the installation was successful.

```python
python -c "import openvino_genai; print(openvino_genai.__version__)"
```

--------------------------------

### Run image-to-image sample

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/image_generation/README.md

Download an initial image and execute the image2image sample to perform denoising based on a provided prompt.

```bash
wget https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/cat.png
./image2image ./dreamlike_anime_1_0_ov/FP16 'cat wizard, gandalf, lord of the rings, detailed, fantasy, cute, adorable, Pixar, Disney, 8k' cat.png
```

--------------------------------

### Define Sample Executable Build Function

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/c/text_generation/CMakeLists.txt

A CMake function to compile C source files, link against the OpenVINO GenAI C library, and configure installation properties.

```cmake
function(add_sample_executable target_name)
    add_executable(${target_name} ${target_name}.c)
    # Specifies that the source file should be compiled as a C source file
    set_source_files_properties(${target_name}.c PROPERTIES LANGUAGE C)
    target_link_libraries(${target_name} PRIVATE openvino::genai::c)
    set_target_properties(${target_name} PROPERTIES
        # Ensure out-of-box LC_RPATH on macOS with SIP
        INSTALL_RPATH_USE_LINK_PATH ON)
    install(TARGETS ${target_name}
            RUNTIME DESTINATION samples_bin/
            COMPONENT samples_bin
            EXCLUDE_FROM_ALL)
endfunction()
```

--------------------------------

### Install VLM Pipeline Executable

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/c/visual_language_chat/CMakeLists.txt

Installs the VLM pipeline executable to a specific runtime destination. This ensures the sample binary is available after installation.

```cmake
install(TARGETS vlm_pipeline_c
        RUNTIME DESTINATION samples_bin/
        COMPONENT samples_bin
        EXCLUDE_FROM_ALL)
```

--------------------------------

### Build OpenVINO GenAI Samples for Windows (PowerShell)

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/c/text_generation/README.md

Set up the environment and build the samples for Windows using PowerShell. Ensure you have sourced the setupvars PowerShell script.

```ps1
.<INSTALL_DIR>\setupvars.ps1
.<INSTALL_DIR>\samples\c\build_samples.ps1
```

--------------------------------

### Configure Sample Executables

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/image_generation/CMakeLists.txt

Standard CMake configuration for building sample executables with OpenVINO GenAI and indicators dependencies.

```cmake
add_executable(stable_diffusion_export_import stable_diffusion_export_import.cpp load_image.cpp imwrite.cpp)

target_include_directories(stable_diffusion_export_import PRIVATE ${CMAKE_BINARY_DIR} "${CMAKE_CURRENT_SOURCE_DIR}")
target_link_libraries(stable_diffusion_export_import PRIVATE openvino::genai indicators::indicators)

set_target_properties(stable_diffusion_export_import PROPERTIES
    # Ensure out of box LC_RPATH on macOS with SIP
    INSTALL_RPATH_USE_LINK_PATH ON)

install(TARGETS stable_diffusion_export_import
        RUNTIME DESTINATION samples_bin/
        COMPONENT samples_bin
        EXCLUDE_FROM_ALL)
```

```cmake
add_executable(inpainting inpainting.cpp load_image.cpp imwrite.cpp)

target_include_directories(inpainting PRIVATE "${CMAKE_CURRENT_SOURCE_DIR}" "${CMAKE_BINARY_DIR}")
target_link_libraries(inpainting PRIVATE openvino::genai indicators::indicators)

set_target_properties(inpainting PROPERTIES
    # Ensure out of box LC_RPATH on macOS with SIP
    INSTALL_RPATH_USE_LINK_PATH ON)

install(TARGETS inpainting
        RUNTIME DESTINATION samples_bin/
        COMPONENT samples_bin
        EXCLUDE_FROM_ALL)
```

--------------------------------

### Build OpenVINO GenAI Samples for Linux/macOS

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/c/text_generation/README.md

Set up the environment and build the samples for Linux and macOS. Ensure you have sourced the setupvars script.

```sh
source <INSTALL_DIR>/setupvars.sh
./<INSTALL_DIR>/samples/c/build_samples.sh
```

--------------------------------

### Install OpenVINO GenAI from Source

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/src/docs/BUILD.md

Install the OpenVINO GenAI package directly from the source code after setting up the environment. This command installs the package into the current Python environment.

```sh
python -m pip install .
```

--------------------------------

### Run Chat Sample

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/js/text_generation/README.md

Launches the interactive chat sample. Provide the path to your model directory as an argument.

```bash
node chat_sample.js model_dir
```

--------------------------------

### Install Export Requirements

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/rag/README.md

Install the necessary dependencies for model conversion.

```sh
pip install --upgrade-strategy eager -r ../../export-requirements.txt
```

--------------------------------

### Set Up Environment using OpenVINO setupvars Script

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/src/js/BUILD.md

Source the OpenVINO setupvars script to configure the environment. This is the recommended method for setting up environment variables.

```sh
source <INSTALL_DIR>/setupvars.sh
```

```cmd
call <INSTALL_DIR>\setupvars.bat
```

```powershell
. <INSTALL_DIR>/setupvars.ps1
```

--------------------------------

### Initialize and Use WhisperPipeline

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/site/docs/use-cases/speech-recognition/_sections/_run_model/_code_example_js.mdx

Demonstrates initializing the WhisperPipeline with a model path and device, then generating text from raw audio data. Ensure 'sample.wav' exists and the modelPath is correctly set.

```javascript
import { WhisperPipeline } from 'openvino-genai-node';
import { readAudio } from './wav_utils.js';

const rawSpeech = readAudio('sample.wav');

const pipeline = await WhisperPipeline(modelPath, "${props.device || 'CPU'}");
const generationConfig = { max_new_tokens: 100 };
const result = await pipeline.generate(rawSpeech, { generationConfig });
console.log(result.texts[0]);

```

--------------------------------

### Install test requirements

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/tests/python_tests/README.md

Install the necessary dependencies for running the Python tests.

```sh
pip install -r tests/python_tests/requirements.txt
```

--------------------------------

### Run Chat Sample

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/text_generation/README.md

Launches the interactive chat sample. Replace <MODEL_DIR> with the path to your model.

```bash
./chat_sample <MODEL_DIR>
```

--------------------------------

### Install ModelScope

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/site/docs/guides/model-preparation/download-openvino-models.mdx

Install the modelscope package to enable model downloads from ModelScope.

```bash
pip install modelscope
```

--------------------------------

### Run LLM Benchmarking Sample

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/text_generation/README.md

Execute the benchmark tool with various options for model path, prompts, and performance metrics.

```bash
./benchmark_genai [OPTIONS]
```

--------------------------------

### Install OpenVINO versions

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/tools/llm_bench/README.md

Install specific or nightly versions of OpenVINO using pip.

```bash
# e.g.
pip install openvino==2024.4.0
# Optional, install the openvino nightly package if needed.
# OpenVINO nightly is pre-release software and has not undergone full release validation or qualification.
pip uninstall openvino
pip install --upgrade --pre openvino openvino-tokenizers --extra-index-url https://storage.openvinotoolkit.org/simple/wheels/nightly
```

--------------------------------

### Install macOS dependencies

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/src/docs/BUILD.md

Install required build tools on macOS using Homebrew.

```sh
brew install coreutils scons
```

--------------------------------

### Install OpenCV Dependency

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/python/video_generation/README.md

Install the specific version of OpenCV required for running the provided samples.

```sh
pip install opencv-python==4.12.0.88
```

--------------------------------

### Build Samples from Archive

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/src/docs/BUILD.md

Configure the environment and build samples after extracting the archive.

```sh
source <INSTALL_DIR>/setupvars.sh
<INSTALL_DIR>/samples/cpp/build_samples.sh
```

```cmd
<INSTALL_DIR>\setupvars.bat
<INSTALL_DIR>\samples\cpp\build_samples_msvc.bat
```

```powershell
.<INSTALL_DIR>\setupvars.ps1
<INSTALL_DIR>\samples\cpp\build_samples.ps1
```

--------------------------------

### Install export requirements

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/js/speech_generation/README.md

Install the necessary dependencies for exporting models using the Optimum CLI.

```sh
pip install --upgrade-strategy eager -r <GENAI_ROOT_DIR>/samples/export-requirements.txt
```

--------------------------------

### Structured Output Example Output

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/js/text_generation/README.md

Example of the output format generated by the structured output sample after processing a prompt.

```text
> Generate a JSON for 2 cars and 1 person with an Irish surname
output:
{"name": "John Doe", "surname": "O'Reilly", "age": 30, "city": "Dublin"}
{"model": "Toyota", "year": 2020, "engine": "hybrid"}
{"model": "Ford", "year": 2019, "color": "red"}
```

--------------------------------

### Apply and Switch LoRA Adapters

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/site/docs/guides/lora-adapters.mdx

Demonstrates initializing an LLM pipeline with multiple LoRA adapters and switching configurations dynamically during generation.

```python
import openvino_genai as ov_genai

# Initialize pipeline with adapters
adapter_config = ov_genai.AdapterConfig()

# Add multiple adapters with different weights
adapter1 = ov_genai.Adapter("path/to/lora1.safetensors")
adapter2 = ov_genai.Adapter("path/to/lora2.safetensors")

adapter_config.add(adapter1, alpha=0.5)
adapter_config.add(adapter2, alpha=0.5)

pipe = ov_genai.LLMPipeline(
    model_path,
    "CPU",
    adapters=adapter_config
)

# Generate with current adapters
output1 = pipe.generate("Generate story about", max_new_tokens=100)

# Switch to different adapter configuration
new_config = ov_genai.AdapterConfig()
new_config.add(adapter1, alpha=1.0)
output2 = pipe.generate(
    "Generate story about",
    max_new_tokens=100,
    adapters=new_config
)
```

```cpp
#include "openvino/genai/llm_pipeline.hpp"

int main() {
    ov::genai::AdapterConfig adapter_config;

    // Add multiple adapters with different weights
    ov::genai::Adapter adapter1("path/to/lora1.safetensors");
    ov::genai::Adapter adapter2("path/to/lora2.safetensors");

    adapter_config.add(adapter1, 0.5f);
    adapter_config.add(adapter2, 0.5f);

    ov::genai::LLMPipeline pipe(
        model_path,
        "CPU",
        ov::genai::adapters(adapter_config)
    );

    // Generate with current adapters
    auto output1 = pipe.generate("Generate story about", ov::genai::max_new_tokens(100));

    // Switch to different adapter configuration
    ov::genai::AdapterConfig new_config;
    new_config.add(adapter1, 1.0f);
    auto output2 = pipe.generate(
        "Generate story about",
        ov::genai::adapters(new_config),
        ov::genai::max_new_tokens(100)
    );
}
```

--------------------------------

### Display benchmark help

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/tools/llm_bench/README.md

Access the help menu for the benchmarking script.

```bash
python ./benchmark.py -h # for more information
```

--------------------------------

### Windows Unicode Encoding Error Example

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/python/video_generation/README.md

Example of the UnicodeEncodeError that may occur when printing output to the Windows console.

```text
UnicodeEncodeError: 'charmap' codec can't encode character '\u25aa' in position 0: character maps to <undefined>
```

--------------------------------

### Set RPATH and Install Targets

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/src/python/CMakeLists.txt

Configures platform-specific RPATHs and defines installation rules for the Python module and distribution files.

```cmake
# setting RPATH / LC_RPATH depending on platform
if(LINUX)
    # to find libopenvino_genai.so in the same folder
    set(py_rpaths "$ORIGIN")
elseif(APPLE)
    # to find libopenvino_genai.dylib in the same folder
    set(py_rpaths "@loader_path")
    if(DEFINED PY_BUILD_CMAKE_PACKAGE_NAME)
        # in case we build pip package, we need to refer to libopenvino.dylib from 'openvino' package
        list(APPEND py_rpaths "@loader_path/../openvino/libs")
    endif()
endif()

if(py_rpaths)
    set_target_properties(${TARGET_NAME} PROPERTIES INSTALL_RPATH "${py_rpaths}")
endif()

install(FILES "${CMAKE_CURRENT_SOURCE_DIR}/openvino_genai/__init__.py"
              "${CMAKE_CURRENT_SOURCE_DIR}/openvino_genai/__init__.pyi"
              "${CMAKE_CURRENT_SOURCE_DIR}/openvino_genai/py_openvino_genai.pyi"
        DESTINATION python/openvino_genai
        COMPONENT pygenai_${Python3_VERSION_MAJOR}_${Python3_VERSION_MINOR})
install(TARGETS ${TARGET_NAME}
        LIBRARY DESTINATION python/openvino_genai
        COMPONENT pygenai_${Python3_VERSION_MAJOR}_${Python3_VERSION_MINOR})

install(FILES "${OpenVINOGenAI_SOURCE_DIR}/LICENSE"
              "${OpenVINOGenAI_SOURCE_DIR}/third-party-programs.txt"
              "${OpenVINOGenAI_SOURCE_DIR}/SECURITY.md"
        DESTINATION "${PY_BUILD_CMAKE_PACKAGE_NAME}-${PY_BUILD_CMAKE_PACKAGE_VERSION}.dist-info"
        COMPONENT wheel_genai
        EXCLUDE_FROM_ALL)

# wheel_genai component is used for wheel generation in pyproject.toml.
# Exclude wheel_genai from normal packaging because there's pygenai_X_Y component for that.
install(TARGETS openvino_genai ${TARGET_NAME}
        LIBRARY DESTINATION openvino_genai COMPONENT wheel_genai EXCLUDE_FROM_ALL
        RUNTIME DESTINATION openvino_genai COMPONENT wheel_genai EXCLUDE_FROM_ALL)
```

--------------------------------

### Register and Build Sample List

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/c/text_generation/CMakeLists.txt

Defines a list of sample targets and iterates through them to invoke the build function.

```cmake
set (SAMPLE_LIST
    greedy_causal_lm_c
    chat_sample_c
    benchmark_genai_c)

foreach(sample IN LISTS SAMPLE_LIST)
    add_sample_executable(${sample})
endforeach()
```

--------------------------------

### Initialize Pipeline and Generate

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/site/docs/concepts/optimization-techniques/visual-token-pruning.md

Create a VLMPipeline instance and execute generation with the provided configuration.

```python
pipe = ov_genai.VLMPipeline(models_path, device, scheduler_config=scheduler_config)
res = pipe.generate(prompt, images=images, generation_config=config)
```

--------------------------------

### Run TaylorSeer Text-to-Video Sample

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/python/video_generation/README.md

Execute the TaylorSeer sample script to compare baseline performance against optimized caching. Requires a model directory and a prompt string.

```bash
python taylorseer_text2video.py model_dir prompt
```

```bash
python taylorseer_text2video.py ./ltx_video_ov/INT8 "a robot dancing in the rain"
```

--------------------------------

### Run Basic Text to Video Generation

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/python/video_generation/README.md

Executes the text2video.py sample to generate a video from a text prompt using a pre-exported OpenVINO model. Specify the model directory and the text prompt as arguments.

```bash
python text2video.py ./ltx_video_ov/FP32 "A woman with long brown hair and light skin smiles at another woman with long blonde hair"
```

--------------------------------

### Install Python Dependencies for JavaScript Tests

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/src/js/BUILD.md

Install the required Python packages for running JavaScript tests. This includes 'openvino-genai' and 'optimum-intel' for model conversion.

```sh
python -m pip install openvino-genai optimum-intel
```

--------------------------------

### Install Dependencies and Export Model

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/whisper_speech_recognition/README.md

Installs necessary dependencies and exports the Whisper model using the optimum-cli. Ensure optimum-intel is upgraded to the latest version.

```sh
pip install --upgrade-strategy eager -r ../../requirements.txt
optimum-cli export openvino --trust-remote-code --model openai/whisper-base whisper-base
```

--------------------------------

### Benchmark GenAI CMake Setup

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/rag/CMakeLists.txt

Includes CMake's FetchContent module and sets a policy for compatibility. This is typically used for managing external dependencies or modules required for benchmarking.

```cmake
include(FetchContent)

if(POLICY CMP0135)
    cmake_policy(SET CMP0135 NEW)
endif()
```

--------------------------------

### Verify Node.js Bindings Installation

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/src/js/BUILD.md

Verify the installation of the OpenVINO GenAI Node.js bindings by checking if the LLMPipeline class can be imported. This confirms the package is accessible in your Node.js environment.

```javascript
node -e "const { LLMPipeline } = require('openvino-genai-node'); console.log(LLMPipeline);"
```

--------------------------------

### Find OpenVINO GenAI Package

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/c/whisper_speech_recognition/CMakeLists.txt

Locates the OpenVINO GenAI package. It searches in the build directory and the OpenVINO installation directory. Ensure OpenVINO is installed or built in one of these locations.

```cmake
find_package(OpenVINOGenAI REQUIRED
    PATHS
        "${CMAKE_BINARY_DIR}"  # Reuse the package from the build.
        ${OpenVINO_DIR}  # GenAI may be installed alogside OpenVINO.
    NO_CMAKE_FIND_ROOT_PATH
)
```

--------------------------------

### Run LLM ReAct Agent Sample

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/js/text_generation/README.md

Launches the LLM ReAct Agent sample for real-time reasoning and action based on user input. Provide the path to your model directory.

```bash
node react_sample.js model_dir
```

--------------------------------

### Run LLM Benchmarking Sample

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/c/text_generation/README.md

Benchmark Large Language Models in OpenVINO GenAI using C. This sample allows customization of the model, prompt, device, and iteration counts.

```sh
./benchmark_genai_c  [-m MODEL] [-p PROMPT] [-nw NUM_WARMUP] [-n NUM_ITER] [-mt MAX_NEW_TOKENS] [-d DEVICE]
```

--------------------------------

### Structured Output - Stage One Example

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/python/text_generation/README.md

Example of a model generating a JSON schema to indicate the number of items of each type requested in a prompt. This is an internal step not shown to the user.

```json
{"person": 1, "car": 2, "transaction": 0}
```

--------------------------------

### Run text-to-speech sample

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/js/speech_generation/README.md

Execute the text-to-speech sample with various configuration options.

```bash
node speech_generation/text2speech.js speecht5_tts "Hello OpenVINO GenAI"
```

```bash
node speech_generation/text2speech.js speecht5_tts "Hello OpenVINO GenAI" --speaker_embedding speaker_embedding.bin
```

```bash
node speech_generation/text2speech.js speecht5_tts "Hello OpenVINO GenAI" GPU
```

```bash
node speech_generation/text2speech.js speecht5_tts "Hello OpenVINO GenAI" --output my_audio.wav
```

--------------------------------

### Run LoRA text-to-image sample

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/image_generation/README.md

Execute the lora_text2image sample with a specified model path, prompt, adapter file, and alpha scaling factor.

```bash
./lora_text2image dreamlike_anime_1_0_ov/FP16 'curly-haired unicorn in the forest, anime, line' soulcard.safetensors 0.7
```

--------------------------------

### Install Dependencies and Export Model

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/visual_language_chat/README.md

Installs necessary Python packages and exports a Hugging Face model to OpenVINO format using the optimum-cli. Ensure optimum-intel is upgraded to the latest version.

```sh
pip install --upgrade-strategy eager -r ../../requirements.txt
optimum-cli export openvino --model openbmb/MiniCPM-V-2_6 --trust-remote-code MiniCPM-V-2_6
```

--------------------------------

### Example Structured JSON Output

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/src/README.md

This is an example of a JSON object that conforms to the `Person` schema defined in the Python code. Note that while the JSON format is guaranteed, the factual correctness or sensibility of the content is not.

```json
{
  "name": "John",
  "surname": "Doe",
  "age": 30,
  "city": "Dublin"
}
```

--------------------------------

### Basic Text to Video Generation Command

Source: https://github.com/openvinotoolkit/openvino.genai/blob/master/samples/cpp/video_generation/README.md

Executes the text-to-video sample to generate a video from a text prompt. Requires the model directory and the prompt as arguments.

```bash
./text2video model_dir prompt
```