diff --git a/README.md b/README.md
index 9542fb78..555180c4 100644
--- a/README.md
+++ b/README.md
@@ -301,6 +301,7 @@ Over 50 models have been optimized/verified on `ipex-llm`, including *LLaMA/LLaM
| Phi-3-vision | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/phi-3-vision) | [link](python/llm/example/GPU/HuggingFace/Multimodal/phi-3-vision) |
| Yuan2 | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/yuan2) | [link](python/llm/example/GPU/HuggingFace/LLM/yuan2) |
| Gemma | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/gemma) | [link](python/llm/example/GPU/HuggingFace/LLM/gemma) |
+| Gemma2 | | [link](python/llm/example/GPU/HuggingFace/LLM/gemma2) |
| DeciLM-7B | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/deciLM-7b) | [link](python/llm/example/GPU/HuggingFace/LLM/deciLM-7b) |
| Deepseek | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/deepseek) | [link](python/llm/example/GPU/HuggingFace/LLM/deepseek) |
| StableLM | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/stablelm) | [link](python/llm/example/GPU/HuggingFace/LLM/stablelm) |
diff --git a/python/llm/example/GPU/HuggingFace/LLM/gemma2/README.md b/python/llm/example/GPU/HuggingFace/LLM/gemma2/README.md
new file mode 100644
index 00000000..c9352357
--- /dev/null
+++ b/python/llm/example/GPU/HuggingFace/LLM/gemma2/README.md
@@ -0,0 +1,144 @@
+# Gemma2
+In this directory, you will find examples on how you could apply IPEX-LLM INT4 optimizations on Google Gemma2 models on [Intel GPUs](../../../README.md). For illustration purposes, we utilize the [google/gemma-2-9b-it](https://huggingface.co/google/gemma-2-9b-it) and [google/gemma-2-2b-it](https://huggingface.co/google/gemma-2-2b-it) as reference Gemma2 models.
+
+## 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.
+
+**Important: According to Gemma2's requirement, please make sure you have installed `transformers==4.43.1` and `trl` to run the example.**
+
+## Example: Predict Tokens using `generate()` API
+In the example [generate.py](./generate.py), we show a basic use case for a Gemma2 model to predict the next N tokens using `generate()` API, with IPEX-LLM INT4 optimizations 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/
+
+# According to Gemma2's requirement, please make sure you are using a stable version of Transformers, 4.43.1 or newer.
+pip install "transformers>=4.43.1"
+pip install trl
+```
+
+#### 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/
+
+# According to Gemma2's requirement, please make sure you are using a stable version of Transformers, 4.43.1 or newer.
+pip install "transformers>=4.43.1"
+pip install trl
+```
+
+### 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
+
+
+For Intel Arc™ A-Series Graphics and Intel Data Center GPU Flex Series
+
+```bash
+export USE_XETLA=OFF
+export SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=1
+export SYCL_CACHE_PERSISTENT=1
+```
+
+
+
+
+
+For Intel Data Center GPU Max Series
+
+```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`.
+
+
+
+
+For Intel iGPU
+
+```bash
+export SYCL_CACHE_PERSISTENT=1
+export BIGDL_LLM_XMX_DISABLED=1
+```
+
+
+
+#### 3.2 Configurations for Windows
+
+
+For Intel iGPU
+
+```cmd
+set SYCL_CACHE_PERSISTENT=1
+set BIGDL_LLM_XMX_DISABLED=1
+```
+
+
+
+
+
+For Intel Arc™ A-Series Graphics
+
+```cmd
+set SYCL_CACHE_PERSISTENT=1
+```
+
+
+
+> [!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
+
+```bash
+python ./generate.py --prompt 'What is AI?'
+```
+
+In the example, several arguments can be passed to satisfy your requirements:
+
+- `--repo-id-or-model-path REPO_ID_OR_MODEL_PATH`: argument defining the huggingface repo id for the Gemma model (e.g. `google/gemma-2-9b-it` and `google/gemma-2-2b-it`) to be downloaded, or the path to the huggingface checkpoint folder. It is default to be `'google/gemma-2-9b-it'`.
+- `--prompt PROMPT`: argument defining the prompt to be infered (with integrated prompt format for chat). It is default to be `'What is AI?'`.
+- `--n-predict N_PREDICT`: argument defining the max number of tokens to predict. It is default to be `32`.
+
+##### Sample Output
+##### [google/gemma-2-9b-it](https://huggingface.co/google/gemma-2-9b-it)
+```log
+Inference time: xxxx s
+-------------------- Output --------------------
+user
+What is AI?
+model
+Artificial intelligence (AI) is a broad field of computer science focused on creating intelligent agents, which are systems that can reason, learn, and act autonomously.
+```
+
+##### [google/gemma-2-2b-it](https://huggingface.co/google/gemma-2-2b-it)
+```log
+Inference time: xxxx s
+-------------------- Output --------------------
+user
+What is AI?
+model
+AI, or Artificial Intelligence, is a broad field of computer science focused on creating intelligent agents, which are systems that can reason, learn, and act like humans
+```
diff --git a/python/llm/example/GPU/HuggingFace/LLM/gemma2/generate.py b/python/llm/example/GPU/HuggingFace/LLM/gemma2/generate.py
new file mode 100644
index 00000000..2e24a2b4
--- /dev/null
+++ b/python/llm/example/GPU/HuggingFace/LLM/gemma2/generate.py
@@ -0,0 +1,81 @@
+#
+# Copyright 2016 The BigDL Authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+import torch
+import time
+import argparse
+
+from ipex_llm.transformers import AutoModelForCausalLM
+from transformers import AutoTokenizer
+
+# The instruction-tuned models use a chat template that must be adhered to for conversational use.
+# see https://huggingface.co/google/gemma-2b-it#chat-template.
+chat = [
+ { "role": "user", "content": "What is AI?" },
+]
+
+if __name__ == '__main__':
+ parser = argparse.ArgumentParser(description='Predict Tokens using `generate()` API for Gemma model')
+ parser.add_argument('--repo-id-or-model-path', type=str, default="google/gemma-2-9b-it",
+ help='The huggingface repo id for the Gemma2 (e.g. `google/gemma-2-9b-it` and `google/gemma-2-2b-it`) to be downloaded'
+ ', or the path to the huggingface checkpoint folder')
+ parser.add_argument('--prompt', type=str, default="What is AI?",
+ help='Prompt to infer')
+ parser.add_argument('--n-predict', type=int, default=32,
+ help='Max tokens to predict')
+
+ args = parser.parse_args()
+ model_path = args.repo_id_or_model_path
+
+ # Load model in 4 bit,
+ # which convert the relevant layers in the model into INT4 format
+ # When running LLMs on Intel iGPUs for Windows users, we recommend setting `cpu_embedding=True` in the from_pretrained function.
+ # This will allow the memory-intensive embedding layer to utilize the CPU instead of iGPU.
+ model = AutoModelForCausalLM.from_pretrained(model_path,
+ load_in_4bit=True,
+ optimize_model=True,
+ trust_remote_code=True,
+ mixed_precision=True,
+ use_cache=True)
+ model = model.to('xpu')
+
+ # Load tokenizer
+ tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
+
+ # Generate predicted tokens
+ with torch.inference_mode():
+ chat[0]['content'] = args.prompt
+ prompt = tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=True)
+ input_ids = tokenizer.encode(prompt, return_tensors="pt").to('xpu')
+ # ipex_llm model needs a warmup, then inference time can be accurate
+ output = model.generate(input_ids,
+ max_new_tokens=args.n_predict)
+
+ # start inference
+ st = time.time()
+ # if your selected model is capable of utilizing previous key/value attentions
+ # to enhance decoding speed, but has `"use_cache": false` in its model config,
+ # it is important to set `use_cache=True` explicitly in the `generate` function
+ # to obtain optimal performance with IPEX-LLM INT4 optimizations
+ output = model.generate(input_ids,
+ max_new_tokens=args.n_predict)
+ torch.xpu.synchronize()
+ end = time.time()
+ output = output.cpu()
+ output_str = tokenizer.decode(output[0], skip_special_tokens=True)
+ print(f'Inference time: {end-st} s')
+ print('-'*20, 'Output', '-'*20)
+ print(output_str)