Add MiniCPM-Llama3-V-2_5 GPU example (#11693)
* Add MiniCPM-Llama3-V-2_5 GPU example * fix
This commit is contained in:
Jin, Qiao
GitHub
parent
808d9a7bae
commit
7f241133da
2 changed files with 219 additions and 0 deletions
|
|
@ -0,0 +1,135 @@
|
|||
# MiniCPM-Llama3-V-2_5
|
||||
In this directory, you will find examples on how you could apply IPEX-LLM INT4 optimizations on MiniCPM-Llama3-V-2_5 models on [Intel GPUs](../../../README.md). For illustration purposes, we utilize the [openbmb/MiniCPM-Llama3-V-2_5](https://huggingface.co/openbmb/MiniCPM-Llama3-V-2_5) as a reference MiniCPM-Llama3-V-2_5 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 `chat()` API
|
||||
In the example [generate.py](./generate.py), we show a basic use case for a MiniCPM-Llama3-V-2_5 model to predict the next N tokens using `chat()` 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/
|
||||
|
||||
pip install transformers==4.41.0 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/
|
||||
|
||||
pip install transformers==4.41.0 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
|
||||
<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 --prompt 'What is in the image?'
|
||||
```
|
||||
|
||||
Arguments info:
|
||||
- `--repo-id-or-model-path REPO_ID_OR_MODEL_PATH`: argument defining the huggingface repo id for the MiniCPM-Llama3-V-2_5 (e.g. `openbmb/MiniCPM-Llama3-V-2_5`) to be downloaded, or the path to the huggingface checkpoint folder. It is default to be `'openbmb/MiniCPM-Llama3-V-2_5'`.
|
||||
- `--image-url-or-path IMAGE_URL_OR_PATH`: argument defining the image to be infered. It is default to be `'http://farm6.staticflickr.com/5268/5602445367_3504763978_z.jpg'`.
|
||||
- `--prompt PROMPT`: argument defining the prompt to be infered (with integrated prompt format for chat). It is default to be `'What is in the image?'`.
|
||||
- `--n-predict N_PREDICT`: argument defining the max number of tokens to predict. It is default to be `32`.
|
||||
|
||||
#### Sample Output
|
||||
|
||||
#### [openbmb/MiniCPM-Llama3-V-2_5](https://huggingface.co/openbmb/MiniCPM-Llama3-V-2_5)
|
||||
|
||||
```log
|
||||
Inference time: xxxx s
|
||||
-------------------- Input --------------------
|
||||
http://farm6.staticflickr.com/5268/5602445367_3504763978_z.jpg
|
||||
-------------------- Prompt --------------------
|
||||
What is in the image?
|
||||
-------------------- Output --------------------
|
||||
The image features a young child holding a white teddy bear. The teddy bear is dressed in a pink outfit. The child appears to be outdoors, with a stone wall and some red flowers in the background.
|
||||
```
|
||||
|
||||
The sample input image is (which is fetched from [COCO dataset](https://cocodataset.org/#explore?id=264959)):
|
||||
|
||||
<a href="http://farm6.staticflickr.com/5268/5602445367_3504763978_z.jpg"><img width=400px src="http://farm6.staticflickr.com/5268/5602445367_3504763978_z.jpg" ></a>
|
||||
|
|
@ -0,0 +1,84 @@
|
|||
#
|
||||
# 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 os
|
||||
import time
|
||||
import argparse
|
||||
import requests
|
||||
from PIL import Image
|
||||
from ipex_llm.transformers import AutoModel
|
||||
from transformers import AutoTokenizer
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
parser = argparse.ArgumentParser(description='Predict Tokens using `generate()` API for openbmb/MiniCPM-Llama3-V-2_5 model')
|
||||
parser.add_argument('--repo-id-or-model-path', type=str, default="openbmb/MiniCPM-Llama3-V-2_5",
|
||||
help='The huggingface repo id for the openbmb/MiniCPM-Llama3-V-2_5 model to be downloaded'
|
||||
', or the path to the huggingface checkpoint folder')
|
||||
parser.add_argument('--image-url-or-path', type=str,
|
||||
default='http://farm6.staticflickr.com/5268/5602445367_3504763978_z.jpg',
|
||||
help='The URL or path to the image to infer')
|
||||
parser.add_argument('--prompt', type=str, default="What is in the image?",
|
||||
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
|
||||
image_path = args.image_url_or_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 = AutoModel.from_pretrained(model_path,
|
||||
load_in_4bit=True,
|
||||
optimize_model=False,
|
||||
trust_remote_code=True,
|
||||
modules_to_not_convert=["vpm", "resampler"],
|
||||
use_cache=True)
|
||||
model = model.float().to(device='xpu')
|
||||
tokenizer = AutoTokenizer.from_pretrained(model_path,
|
||||
trust_remote_code=True)
|
||||
model.eval()
|
||||
|
||||
query = args.prompt
|
||||
if os.path.exists(image_path):
|
||||
image = Image.open(image_path).convert('RGB')
|
||||
else:
|
||||
image = Image.open(requests.get(image_path, stream=True).raw).convert('RGB')
|
||||
|
||||
# Generate predicted tokens
|
||||
# here the prompt tuning refers to https://huggingface.co/openbmb/MiniCPM-Llama3-V-2_5/blob/main/README.md
|
||||
msgs = [{'role': 'user', 'content': args.prompt}]
|
||||
st = time.time()
|
||||
res = model.chat(
|
||||
image=image,
|
||||
msgs=msgs,
|
||||
context=None,
|
||||
tokenizer=tokenizer,
|
||||
sampling=False,
|
||||
temperature=0.7
|
||||
)
|
||||
end = time.time()
|
||||
print(f'Inference time: {end-st} s')
|
||||
print('-'*20, 'Input', '-'*20)
|
||||
print(image_path)
|
||||
print('-'*20, 'Prompt', '-'*20)
|
||||
print(args.prompt)
|
||||
output_str = res
|
||||
print('-'*20, 'Output', '-'*20)
|
||||
print(output_str)
|
||||
Loading…
Reference in a new issue