# Llama3.2-Vision
In this directory, you will find examples on how you could use IPEX-LLM `optimize_model` API to accelerate Llama3.2-Vision models on [Intel GPUs](../../../README.md). For illustration purposes, we utilize the [meta-llama/Llama-3.2-11B-Vision-Instruct](https://huggingface.co/meta-llama/Llama-3.2-11B-Vision-Instruct) as a reference Llama3.2-Vision model.

## 0. Requirements
To run these examples with IPEX-LLM on Intel GPUs, we have some recommended requirements for your machine, please refer to [here](../../../README.md#requirements) for more information.

## Example: Predict Tokens using `generate()` API
In the example [generate.py](./generate.py), we show a basic use case for a Llama3.2-Vision model to predict the next N tokens using `generate()` API, with IPEX-LLM 'optimize_model' API on Intel GPUs.
### 1. Install
#### 1.1 Installation on Linux
We suggest using conda to manage environment:
```bash
conda create -n llm python=3.11
conda activate llm
# below command will install intel_extension_for_pytorch==2.1.10+xpu as default
pip install --pre --upgrade ipex-llm[xpu] --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/

pip install transformers==4.45.0
```

#### 1.2 Installation on Windows
We suggest using conda to manage environment:
```bash
conda create -n llm python=3.11 libuv
conda activate llm

# below command will install intel_extension_for_pytorch==2.1.10+xpu as default
pip install --pre --upgrade ipex-llm[xpu] --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/

pip install transformers==4.45.0
```

### 2. Configures OneAPI environment variables for Linux

> [!NOTE]
> Skip this step if you are running on Windows.

This is a required step on Linux for APT or offline installed oneAPI. Skip this step for PIP-installed oneAPI.

```bash
source /opt/intel/oneapi/setvars.sh
```

### 3. Runtime Configurations
For optimal performance, it is recommended to set several environment variables. Please check out the suggestions based on your device.
#### 3.1 Configurations for Linux
<details>

<summary>For Intel Arc™ A-Series Graphics and Intel Data Center GPU Flex Series</summary>

```bash
export USE_XETLA=OFF
export SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=1
export SYCL_CACHE_PERSISTENT=1
```

</details>

<details>

<summary>For Intel Data Center GPU Max Series</summary>

```bash
export LD_PRELOAD=${LD_PRELOAD}:${CONDA_PREFIX}/lib/libtcmalloc.so
export SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=1
export SYCL_CACHE_PERSISTENT=1
export ENABLE_SDP_FUSION=1
```
> Note: Please note that `libtcmalloc.so` can be installed by `conda install -c conda-forge -y gperftools=2.10`.
</details>

<details>

<summary>For Intel iGPU</summary>

```bash
export SYCL_CACHE_PERSISTENT=1
export BIGDL_LLM_XMX_DISABLED=1
```

</details>

#### 3.2 Configurations for Windows
<details>

<summary>For Intel iGPU</summary>

```cmd
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
```

</details>

<details>

<summary>For Intel Arc™ A-Series Graphics</summary>

```cmd
set SYCL_CACHE_PERSISTENT=1
```

</details>

> [!NOTE]
> For the first time that each model runs on Intel iGPU/Intel Arc™ A300-Series or Pro A60, it may take several minutes to compile.
### 4. Running examples

```
python ./generate.py
```

Arguments info:
- `--repo-id-or-model-path REPO_ID_OR_MODEL_PATH`: argument defining the huggingface repo id for the Llama3.2-Vision model (e.g. `meta-llama/Llama-3.2-11B-Vision-Instruct`) to be downloaded, or the path to the huggingface checkpoint folder. It is default to be `'meta-llama/Llama-3.2-11B-Vision-Instruct'`.
- `--image-url-or-path IMAGE_URL_OR_PATH`: argument defining the image to be infered. It is default to be `'https://hf-mirror.com/datasets/huggingface/documentation-images/resolve/0052a70beed5bf71b92610a43a52df6d286cd5f3/diffusers/rabbit.jpg'`.
- `--prompt PROMPT`: argument defining the prompt to be infered (with integrated prompt format for chat). It is default to be `'Describe image in detail'`.
- `--n-predict N_PREDICT`: argument defining the max number of tokens to predict. It is default to be `32`.

#### Sample Output
#### [meta-llama/Llama-3.2-11B-Vision-Instruct](https://huggingface.co/meta-llama/Llama-3.2-11B-Vision-Instruct)

```log
Inference time: xxxx s
-------------------- Prompt --------------------
Describe image in detail
-------------------- Output --------------------
This image features a charming anthropomorphic rabbit standing on a dirt path, surrounded by a picturesque rural landscape.

The rabbit, with its light brown fur and distinctive large
```

The sample input image is:

<a href="https://hf-mirror.com/datasets/huggingface/documentation-images/resolve/0052a70beed5bf71b92610a43a52df6d286cd5f3/diffusers/rabbit.jpg"><img width=400px src="https://hf-mirror.com/datasets/huggingface/documentation-images/resolve/0052a70beed5bf71b92610a43a52df6d286cd5f3/diffusers/rabbit.jpg" ></a>