Add qwen2 example (#11252)
* Add GPU example for Qwen2 * Update comments in README * Update README for Qwen2 GPU example * Add CPU example for Qwen2 Sample Output under README pending * Update generate.py and README for CPU Qwen2 * Update GPU example for Qwen2 * Small update * Small fix * Add Qwen2 table * Update README for Qwen2 CPU and GPU Update sample output under README --------- Co-authored-by: Zijie Li <michael20001122@gmail.com>
This commit is contained in:
Yuwen Hu
GitHub
parent
85df5e7699
commit
8c36b5bdde
10 changed files with 788 additions and 0 deletions
|
|
@ -169,6 +169,7 @@ Over 50 models have been optimized/verified on `ipex-llm`, including *LLaMA/LLaM
|
|||
| InternLM | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/internlm) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/internlm) |
|
||||
| Qwen | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/qwen) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/qwen) |
|
||||
| Qwen1.5 | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/qwen1.5) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/qwen1.5) |
|
||||
| Qwen2 | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/qwen2) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/qwe2) |
|
||||
| Qwen-VL | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/qwen-vl) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/qwen-vl) |
|
||||
| Aquila | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/aquila) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/aquila) |
|
||||
| Aquila2 | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/aquila2) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/aquila2) |
|
||||
|
|
|
|||
|
|
@ -363,6 +363,13 @@ Verified Models
|
|||
<td>
|
||||
<a href="https://github.com/intel-analytics/ipex-llm/tree/main/python/llm/example/GPU/HF-Transformers-AutoModels/Model/qwen1.5">link</a></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td>Qwen2</td>
|
||||
<td>
|
||||
<a href="https://github.com/intel-analytics/ipex-llm/tree/main/python/llm/example/CPU/HF-Transformers-AutoModels/Model/qwen2">link</a></td>
|
||||
<td>
|
||||
<a href="https://github.com/intel-analytics/ipex-llm/tree/main/python/llm/example/GPU/HF-Transformers-AutoModels/Model/qwen2">link</a></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td>Qwen-VL</td>
|
||||
<td>
|
||||
|
|
|
|||
|
|
@ -0,0 +1,83 @@
|
|||
# Qwen2
|
||||
|
||||
In this directory, you will find examples on how you could apply IPEX-LLM INT4 optimizations on Qwen2 models. For illustration purposes, we utilize the [Qwen/Qwen2-7B-Instruct](https://huggingface.co/Qwen/Qwen2-7B-Instruct) as a reference Qwen2 model.
|
||||
|
||||
## 0. Requirements
|
||||
To run these examples with IPEX-LLM, we have some recommended requirements for your machine, please refer to [here](../README.md#recommended-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 Qwen model to predict the next N tokens using `generate()` API, with IPEX-LLM INT4 optimizations.
|
||||
### 1. Install
|
||||
We suggest using conda to manage environment:
|
||||
|
||||
On Linux:
|
||||
|
||||
```bash
|
||||
conda create -n llm python=3.11
|
||||
conda activate llm
|
||||
|
||||
# install ipex-llm with 'all' option
|
||||
pip install --pre --upgrade ipex-llm[all] --extra-index-url https://download.pytorch.org/whl/cpu
|
||||
pip install transformers==4.37.0 # install the transformers which support Qwen2
|
||||
```
|
||||
|
||||
On Windows:
|
||||
|
||||
```cmd
|
||||
conda create -n llm python=3.11
|
||||
conda activate llm
|
||||
|
||||
pip install --pre --upgrade ipex-llm[all]
|
||||
pip install transformers==4.37.0
|
||||
```
|
||||
|
||||
### 2. Run
|
||||
```
|
||||
python ./generate.py --repo-id-or-model-path REPO_ID_OR_MODEL_PATH --prompt PROMPT --n-predict N_PREDICT
|
||||
```
|
||||
|
||||
Arguments info:
|
||||
- `--repo-id-or-model-path REPO_ID_OR_MODEL_PATH`: argument defining the huggingface repo id for the Qwen model to be downloaded, or the path to the huggingface checkpoint folder. It is default to be `'Qwen/Qwen2-7B-Instruct'`.
|
||||
- `--prompt PROMPT`: argument defining the prompt to be infered (with integrated prompt format for chat). It is default to be `'AI是什么?'`.
|
||||
- `--n-predict N_PREDICT`: argument defining the max number of tokens to predict. It is default to be `32`.
|
||||
|
||||
> **Note**: When loading the model in 4-bit, IPEX-LLM converts linear layers in the model into INT4 format. In theory, a *X*B model saved in 16-bit will requires approximately 2*X* GB of memory for loading, and ~0.5*X* GB memory for further inference.
|
||||
>
|
||||
> Please select the appropriate size of the Qwen model based on the capabilities of your machine.
|
||||
|
||||
#### 2.1 Client
|
||||
On client Windows machine, it is recommended to run directly with full utilization of all cores:
|
||||
```cmd
|
||||
python ./generate.py
|
||||
```
|
||||
|
||||
#### 2.2 Server
|
||||
For optimal performance on server, it is recommended to set several environment variables (refer to [here](../README.md#best-known-configuration-on-linux) for more information), and run the example with all the physical cores of a single socket.
|
||||
|
||||
E.g. on Linux,
|
||||
```bash
|
||||
# set IPEX-LLM env variables
|
||||
source ipex-llm-init
|
||||
|
||||
# e.g. for a server with 48 cores per socket
|
||||
export OMP_NUM_THREADS=48
|
||||
numactl -C 0-47 -m 0 python ./generate.py
|
||||
```
|
||||
|
||||
#### 2.3 Sample Output
|
||||
##### [Qwen/Qwen2-7B-Instruct](https://huggingface.co/Qwen/Qwen2-7B-Instruct)
|
||||
```log
|
||||
Inference time: xxxx s
|
||||
-------------------- Prompt --------------------
|
||||
AI是什么？
|
||||
-------------------- Output --------------------
|
||||
AI，即人工智能（Artificial Intelligence），是一种计算机科学领域，旨在开发能够模拟、延伸和增强人类智能的算法和系统。人工 智能涉及许多
|
||||
```
|
||||
|
||||
```log
|
||||
Inference time: xxxx s
|
||||
-------------------- Prompt --------------------
|
||||
What is AI?
|
||||
-------------------- Output --------------------
|
||||
AI stands for Artificial Intelligence, which is the simulation of human intelligence in machines that are programmed to think and learn like humans and mimic their actions. The term may
|
||||
```
|
||||
|
|
@ -0,0 +1,80 @@
|
|||
#
|
||||
# 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
|
||||
import numpy as np
|
||||
|
||||
from transformers import AutoTokenizer
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
parser = argparse.ArgumentParser(description='Qwen2-7B-Instruct')
|
||||
parser.add_argument('--repo-id-or-model-path', type=str, default="Qwen/Qwen2-7B-Instruct",
|
||||
help='The huggingface repo id for the Qwen2 model to be downloaded'
|
||||
', or the path to the huggingface checkpoint folder')
|
||||
parser.add_argument('--prompt', type=str, default="AI是什么？",
|
||||
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
|
||||
|
||||
|
||||
from ipex_llm.transformers import AutoModelForCausalLM
|
||||
model = AutoModelForCausalLM.from_pretrained(model_path,
|
||||
load_in_4bit=True,
|
||||
optimize_model=True,
|
||||
trust_remote_code=True,
|
||||
use_cache=True)
|
||||
|
||||
# Load tokenizer
|
||||
tokenizer = AutoTokenizer.from_pretrained(model_path,
|
||||
trust_remote_code=True)
|
||||
|
||||
prompt = args.prompt
|
||||
|
||||
# Generate predicted tokens
|
||||
with torch.inference_mode():
|
||||
# The following code for generation is adapted from https://huggingface.co/Qwen/Qwen2-7B-Instruct#quickstart
|
||||
messages = [
|
||||
{"role": "system", "content": "You are a helpful assistant."},
|
||||
{"role": "user", "content": prompt}
|
||||
]
|
||||
text = tokenizer.apply_chat_template(
|
||||
messages,
|
||||
tokenize=False,
|
||||
add_generation_prompt=True
|
||||
)
|
||||
model_inputs = tokenizer([text], return_tensors="pt")
|
||||
st = time.time()
|
||||
generated_ids = model.generate(
|
||||
model_inputs.input_ids,
|
||||
max_new_tokens=args.n_predict
|
||||
)
|
||||
end = time.time()
|
||||
generated_ids = [
|
||||
output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
|
||||
]
|
||||
|
||||
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
|
||||
print(f'Inference time: {end-st} s')
|
||||
print('-'*20, 'Prompt', '-'*20)
|
||||
print(prompt)
|
||||
print('-'*20, 'Output', '-'*20)
|
||||
print(response)
|
||||
84
python/llm/example/CPU/PyTorch-Models/Model/qwen2/README.md
Normal file
84
python/llm/example/CPU/PyTorch-Models/Model/qwen2/README.md
Normal file
|
|
@ -0,0 +1,84 @@
|
|||
# Qwen2
|
||||
In this directory, you will find examples on how you could use IPEX-LLM `optimize_model` API to accelerate Qwen2 models. For illustration purposes, we utilize the [Qwen/Qwen2-7B-Instruct](https://huggingface.co/Qwen/Qwen2-7B-Instruct) as reference Qwen2 model.
|
||||
|
||||
## Requirements
|
||||
To run these examples with IPEX-LLM, we have some recommended requirements for your machine, please refer to [here](../README.md#recommended-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 Qwen2 model to predict the next N tokens using `generate()` API, with IPEX-LLM INT4 optimizations.
|
||||
### 1. Install
|
||||
We suggest using conda to manage the Python environment. For more information about conda installation, please refer to [here](https://conda-forge.org/download/).
|
||||
|
||||
After installing conda, create a Python environment for IPEX-LLM:
|
||||
|
||||
On Linux:
|
||||
|
||||
```bash
|
||||
conda create -n llm python=3.11 # recommend to use Python 3.11
|
||||
conda activate llm
|
||||
|
||||
# install the latest ipex-llm nightly build with 'all' option
|
||||
pip install --pre --upgrade ipex-llm[all] --extra-index-url https://download.pytorch.org/whl/cpu
|
||||
pip install transformers==4.37.0 # install transformers which supports Qwen2
|
||||
```
|
||||
|
||||
On Windows:
|
||||
|
||||
```cmd
|
||||
conda create -n llm python=3.11
|
||||
conda activate llm
|
||||
|
||||
pip install --pre --upgrade ipex-llm[all]
|
||||
pip install transformers==4.37.0
|
||||
```
|
||||
|
||||
### 2. Run
|
||||
```
|
||||
python ./generate.py --repo-id-or-model-path REPO_ID_OR_MODEL_PATH --prompt PROMPT --n-predict N_PREDICT
|
||||
```
|
||||
|
||||
Arguments info:
|
||||
- `--repo-id-or-model-path REPO_ID_OR_MODEL_PATH`: argument defining the huggingface repo id for the Qwen2 to be downloaded, or the path to the huggingface checkpoint folder. It is default to be `'Qwen/Qwen2-7B-Instruct'`.
|
||||
- `--prompt PROMPT`: argument defining the prompt to be infered (with integrated prompt format for chat). It is default to be `'AI是什么？'`.
|
||||
- `--n-predict N_PREDICT`: argument defining the max number of tokens to predict. It is default to be `32`.
|
||||
|
||||
> **Note**: When loading the model in 4-bit, IPEX-LLM converts linear layers in the model into INT4 format. In theory, a *X*B model saved in 16-bit will requires approximately 2*X* GB of memory for loading, and ~0.5*X* GB memory for further inference.
|
||||
>
|
||||
> Please select the appropriate size of the Qwen model based on the capabilities of your machine.
|
||||
|
||||
#### 2.1 Client
|
||||
On client Windows machines, it is recommended to run directly with full utilization of all cores:
|
||||
```cmd
|
||||
python ./generate.py --prompt 'What is AI?'
|
||||
```
|
||||
|
||||
#### 2.2 Server
|
||||
For optimal performance on server, it is recommended to set several environment variables (refer to [here](../README.md#best-known-configuration-on-linux) for more information), and run the example with all the physical cores of a single socket.
|
||||
|
||||
E.g. on Linux,
|
||||
```bash
|
||||
# set IPEX-LLM env variables
|
||||
source ipex-llm-init
|
||||
|
||||
# e.g. for a server with 48 cores per socket
|
||||
export OMP_NUM_THREADS=48
|
||||
numactl -C 0-47 -m 0 python ./generate.py --prompt 'What is AI?'
|
||||
```
|
||||
|
||||
#### 2.3 Sample Output
|
||||
##### [Qwen/Qwen2-7B-Instruct](https://huggingface.co/Qwen/Qwen2-7B-Instruct)
|
||||
```log
|
||||
Inference time: xxxx s
|
||||
-------------------- Prompt --------------------
|
||||
AI是什么？
|
||||
-------------------- Output --------------------
|
||||
AI，即人工智能（Artificial Intelligence），是一门研究、开发用于模拟、延伸和扩展人的智能的理论、方法、技术及应用系统的学科
|
||||
```
|
||||
|
||||
```log
|
||||
Inference time: xxxx s
|
||||
-------------------- Prompt --------------------
|
||||
What is AI?
|
||||
-------------------- Output --------------------
|
||||
AI stands for Artificial Intelligence, which refers to the development of computer systems that can perform tasks that typically require human intelligence. These tasks may include learning from experience,
|
||||
```
|
||||
|
|
@ -0,0 +1,82 @@
|
|||
#
|
||||
# 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
|
||||
import numpy as np
|
||||
|
||||
from transformers import AutoTokenizer
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
parser = argparse.ArgumentParser(description='Qwen2-7B-Instruct')
|
||||
parser.add_argument('--repo-id-or-model-path', type=str, default="Qwen/Qwen2-7B-Instruct",
|
||||
help='The huggingface repo id for the Qwen2 model to be downloaded'
|
||||
', or the path to the huggingface checkpoint folder')
|
||||
parser.add_argument('--prompt', type=str, default="AI是什么？",
|
||||
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
|
||||
|
||||
|
||||
from transformers import AutoModelForCausalLM
|
||||
model = AutoModelForCausalLM.from_pretrained(model_path,
|
||||
trust_remote_code=True,
|
||||
torch_dtype='auto',
|
||||
low_cpu_mem_usage=True,
|
||||
use_cache=True)
|
||||
|
||||
|
||||
# Load tokenizer
|
||||
tokenizer = AutoTokenizer.from_pretrained(model_path,
|
||||
trust_remote_code=True)
|
||||
from ipex_llm import optimize_model
|
||||
model = optimize_model(model)
|
||||
|
||||
prompt = args.prompt
|
||||
# Generate predicted tokens
|
||||
with torch.inference_mode():
|
||||
# The following code for generation is adapted from https://huggingface.co/Qwen/Qwen2-7B-Instruct#quickstart
|
||||
messages = [
|
||||
{"role": "system", "content": "You are a helpful assistant."},
|
||||
{"role": "user", "content": prompt}
|
||||
]
|
||||
text = tokenizer.apply_chat_template(
|
||||
messages,
|
||||
tokenize=False,
|
||||
add_generation_prompt=True
|
||||
)
|
||||
model_inputs = tokenizer([text], return_tensors="pt")
|
||||
st = time.time()
|
||||
generated_ids = model.generate(
|
||||
model_inputs.input_ids,
|
||||
max_new_tokens=args.n_predict
|
||||
)
|
||||
end = time.time()
|
||||
generated_ids = [
|
||||
output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
|
||||
]
|
||||
|
||||
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
|
||||
print(f'Inference time: {end-st} s')
|
||||
print('-'*20, 'Prompt', '-'*20)
|
||||
print(prompt)
|
||||
print('-'*20, 'Output', '-'*20)
|
||||
print(response)
|
||||
|
|
@ -0,0 +1,134 @@
|
|||
# Qwen2
|
||||
In this directory, you will find examples on how you could apply IPEX-LLM INT4 optimizations on Qwen2 models on [Intel GPUs](../../../README.md). For illustration purposes, we utilize the [Qwen/Qwen2-7B-Instruct](https://huggingface.co/Qwen/Qwen2-7B-Instruct) as a reference InternLM 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 Qwen2 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/
|
||||
|
||||
pip install transformers==4.37.0 # install transformers which supports Qwen2
|
||||
```
|
||||
|
||||
#### 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==4.37.0 # install transformers which supports Qwen2
|
||||
```
|
||||
|
||||
### 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 --repo-id-or-model-path REPO_ID_OR_MODEL_PATH --prompt PROMPT --n-predict N_PREDICT
|
||||
```
|
||||
|
||||
Arguments info:
|
||||
- `--repo-id-or-model-path REPO_ID_OR_MODEL_PATH`: argument defining the huggingface repo id for the Qwen2 model (e.g. `Qwen/Qwen2-7B-Instruct`) to be downloaded, or the path to the huggingface checkpoint folder. It is default to be `'Qwen/Qwen2-7B-Instruct'`.
|
||||
- `--prompt PROMPT`: argument defining the prompt to be infered (with integrated prompt format for chat). It is default to be `'AI是什么？'`.
|
||||
- `--n-predict N_PREDICT`: argument defining the max number of tokens to predict. It is default to be `32`.
|
||||
|
||||
#### Sample Output
|
||||
##### [Qwen/Qwen2-7B-Instruct](https://huggingface.co/Qwen/Qwen2-7B-Instruct)
|
||||
```log
|
||||
Inference time: xxxx s
|
||||
-------------------- Prompt --------------------
|
||||
AI是什么？
|
||||
-------------------- Output --------------------
|
||||
AI，即人工智能（Artificial Intelligence），是一种模拟人类智能行为的技术。它使计算机系统能够执行类似于人类的思维活动，例 如学习、推理
|
||||
```
|
||||
|
||||
```log
|
||||
Inference time: xxxx s
|
||||
-------------------- Prompt --------------------
|
||||
What is AI?
|
||||
-------------------- Output --------------------
|
||||
AI, or Artificial Intelligence, refers to the simulation of human intelligence in machines that are programmed to think and learn like humans and mimic their actions. The term may
|
||||
```
|
||||
|
|
@ -0,0 +1,92 @@
|
|||
#
|
||||
# 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 transformers import AutoTokenizer
|
||||
from ipex_llm import optimize_model
|
||||
import numpy as np
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
parser = argparse.ArgumentParser(description='Qwen2-7B-Instruct')
|
||||
parser.add_argument('--repo-id-or-model-path', type=str, default="Qwen/Qwen2-7B-Instruct",
|
||||
help='The huggingface repo id for the Qwen2 model to be downloaded'
|
||||
', or the path to the huggingface checkpoint folder')
|
||||
parser.add_argument('--prompt', type=str, default="AI是什么？",
|
||||
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
|
||||
|
||||
|
||||
from ipex_llm.transformers import AutoModelForCausalLM
|
||||
# Load model in 4 bit,
|
||||
# which convert the relevant layers in the model into INT4 format
|
||||
model = AutoModelForCausalLM.from_pretrained(model_path,
|
||||
load_in_4bit=True,
|
||||
optimize_model=True,
|
||||
trust_remote_code=True,
|
||||
use_cache=True)
|
||||
model = model.to("xpu")
|
||||
|
||||
# Load tokenizer
|
||||
tokenizer = AutoTokenizer.from_pretrained(model_path,
|
||||
trust_remote_code=True)
|
||||
|
||||
prompt = args.prompt
|
||||
|
||||
# Generate predicted tokens
|
||||
with torch.inference_mode():
|
||||
# The following code for generation is adapted from https://huggingface.co/Qwen/Qwen2-7B-Instruct#quickstart
|
||||
messages = [
|
||||
{"role": "system", "content": "You are a helpful assistant."},
|
||||
{"role": "user", "content": prompt}
|
||||
]
|
||||
text = tokenizer.apply_chat_template(
|
||||
messages,
|
||||
tokenize=False,
|
||||
add_generation_prompt=True
|
||||
)
|
||||
model_inputs = tokenizer([text], return_tensors="pt").to("xpu")
|
||||
# warmup
|
||||
generated_ids = model.generate(
|
||||
model_inputs.input_ids,
|
||||
max_new_tokens=args.n_predict
|
||||
)
|
||||
|
||||
st = time.time()
|
||||
generated_ids = model.generate(
|
||||
model_inputs.input_ids,
|
||||
max_new_tokens=args.n_predict
|
||||
)
|
||||
torch.xpu.synchronize()
|
||||
end = time.time()
|
||||
generated_ids = generated_ids.cpu()
|
||||
generated_ids = [
|
||||
output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
|
||||
]
|
||||
|
||||
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
|
||||
print(f'Inference time: {end-st} s')
|
||||
print('-'*20, 'Prompt', '-'*20)
|
||||
print(prompt)
|
||||
print('-'*20, 'Output', '-'*20)
|
||||
print(response)
|
||||
134
python/llm/example/GPU/PyTorch-Models/Model/qwen2/README.md
Normal file
134
python/llm/example/GPU/PyTorch-Models/Model/qwen2/README.md
Normal file
|
|
@ -0,0 +1,134 @@
|
|||
# Qwen2
|
||||
In this directory, you will find examples on how you could use IPEX-LLM `optimize_model` API to accelerate Qwen2 models on [Intel GPUs](../../../README.md). For illustration purposes, we utilize the [Qwen/Qwen2-7B-Instruct](https://huggingface.co/Qwen/Qwen2-7B-Instruct) as a reference InternLM 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 Qwen2 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/
|
||||
|
||||
pip install transformers==4.37.0 # install transformers which supports Qwen2
|
||||
```
|
||||
|
||||
#### 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==4.37.0 # install transformers which supports Qwen2
|
||||
```
|
||||
|
||||
### 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 --repo-id-or-model-path REPO_ID_OR_MODEL_PATH --prompt PROMPT --n-predict N_PREDICT
|
||||
```
|
||||
|
||||
Arguments info:
|
||||
- `--repo-id-or-model-path REPO_ID_OR_MODEL_PATH`: argument defining the huggingface repo id for the Qwen2 model (e.g. `Qwen/Qwen2-7B-Instruct`) to be downloaded, or the path to the huggingface checkpoint folder. It is default to be `'Qwen/Qwen2-7B-Instruct'`.
|
||||
- `--prompt PROMPT`: argument defining the prompt to be infered (with integrated prompt format for chat). It is default to be `'AI是什么？'`.
|
||||
- `--n-predict N_PREDICT`: argument defining the max number of tokens to predict. It is default to be `32`.
|
||||
|
||||
#### Sample Output
|
||||
##### [Qwen/Qwen2-7B-Instruct](https://huggingface.co/Qwen/Qwen2-7B-Instruct)
|
||||
```log
|
||||
Inference time: xxxx s
|
||||
-------------------- Prompt --------------------
|
||||
AI是什么？
|
||||
-------------------- Output --------------------
|
||||
AI是人工智能（Artificial Intelligence）的缩写。它指的是通过计算机程序、算法和模型来模拟、延伸和扩展人类智能的一门学科
|
||||
```
|
||||
|
||||
```log
|
||||
Inference time: xxxx s
|
||||
-------------------- Prompt --------------------
|
||||
What is AI?
|
||||
-------------------- Output --------------------
|
||||
AI stands for Artificial Intelligence. It refers to the simulation of human intelligence in machines that are programmed to think and work like humans. This includes learning from experience,
|
||||
```
|
||||
|
|
@ -0,0 +1,91 @@
|
|||
#
|
||||
# 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 transformers import AutoTokenizer
|
||||
from ipex_llm import optimize_model
|
||||
import numpy as np
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
parser = argparse.ArgumentParser(description='Qwen2-7B-Instruct')
|
||||
parser.add_argument('--repo-id-or-model-path', type=str, default="Qwen/Qwen2-7B-Instruct",
|
||||
help='The huggingface repo id for the Qwen2 model to be downloaded'
|
||||
', or the path to the huggingface checkpoint folder')
|
||||
parser.add_argument('--prompt', type=str, default="AI是什么？",
|
||||
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
|
||||
|
||||
|
||||
from transformers import AutoModelForCausalLM
|
||||
from ipex_llm import optimize_model
|
||||
model = AutoModelForCausalLM.from_pretrained(model_path,
|
||||
trust_remote_code=True,
|
||||
torch_dtype = 'auto',
|
||||
low_cpu_mem_usage=True,
|
||||
use_cache=True)
|
||||
model = optimize_model(model)
|
||||
model = model.to("xpu")
|
||||
|
||||
# Load tokenizer
|
||||
tokenizer = AutoTokenizer.from_pretrained(model_path,
|
||||
trust_remote_code=True)
|
||||
|
||||
prompt = args.prompt
|
||||
# Generate predicted tokens
|
||||
with torch.inference_mode():
|
||||
# The following code for generation is adapted from https://huggingface.co/Qwen/Qwen2-7B-Instruct#quickstart
|
||||
messages = [
|
||||
{"role": "system", "content": "You are a helpful assistant."},
|
||||
{"role": "user", "content": prompt}
|
||||
]
|
||||
text = tokenizer.apply_chat_template(
|
||||
messages,
|
||||
tokenize=False,
|
||||
add_generation_prompt=True
|
||||
)
|
||||
model_inputs = tokenizer([text], return_tensors="pt").to("xpu")
|
||||
# warmup
|
||||
generated_ids = model.generate(
|
||||
model_inputs.input_ids,
|
||||
max_new_tokens=args.n_predict
|
||||
)
|
||||
|
||||
st = time.time()
|
||||
generated_ids = model.generate(
|
||||
model_inputs.input_ids,
|
||||
max_new_tokens=args.n_predict
|
||||
)
|
||||
torch.xpu.synchronize()
|
||||
end = time.time()
|
||||
generated_ids = generated_ids.cpu()
|
||||
generated_ids = [
|
||||
output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
|
||||
]
|
||||
|
||||
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
|
||||
print(f'Inference time: {end-st} s')
|
||||
print('-'*20, 'Prompt', '-'*20)
|
||||
print(prompt)
|
||||
print('-'*20, 'Output', '-'*20)
|
||||
print(response)
|
||||
Loading…
Reference in a new issue