Add Qwen1.5-7B-Chat (#10113)
* add Qwen1.5-7B-Chat * modify Qwen1.5 example * update README * update prompt format * update folder name and example README * add Chinese prompt sample output * update link in README * correct the link * update transformer version
This commit is contained in:
Zhicun
GitHub
parent
56ad781f2f
commit
c7e839e66c
10 changed files with 797 additions and 0 deletions
|
|
@ -186,6 +186,7 @@ Over 40 models have been optimized/verified on `bigdl-llm`, including *LLaMA/LLa
|
||||||
| Bark | [link](python/llm/example/CPU/PyTorch-Models/Model/bark) | [link](python/llm/example/GPU/PyTorch-Models/Model/bark) |
|
| Bark | [link](python/llm/example/CPU/PyTorch-Models/Model/bark) | [link](python/llm/example/GPU/PyTorch-Models/Model/bark) |
|
||||||
| SpeechT5 | | [link](python/llm/example/GPU/PyTorch-Models/Model/speech-t5) |
|
| SpeechT5 | | [link](python/llm/example/GPU/PyTorch-Models/Model/speech-t5) |
|
||||||
| Ziya-Coding-34B-v1.0 | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/ziya) | |
|
| Ziya-Coding-34B-v1.0 | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/ziya) | |
|
||||||
|
| Qwen1.5 | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/qwen1.5) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/qwen1.5) |
|
||||||
|
|
||||||
***For more details, please refer to the `bigdl-llm` [Document](https://test-bigdl-llm.readthedocs.io/en/main/doc/LLM/index.html), [Readme](python/llm), [Tutorial](https://github.com/intel-analytics/bigdl-llm-tutorial) and [API Doc](https://bigdl.readthedocs.io/en/latest/doc/PythonAPI/LLM/index.html).***
|
***For more details, please refer to the `bigdl-llm` [Document](https://test-bigdl-llm.readthedocs.io/en/main/doc/LLM/index.html), [Readme](python/llm), [Tutorial](https://github.com/intel-analytics/bigdl-llm-tutorial) and [API Doc](https://bigdl.readthedocs.io/en/latest/doc/PythonAPI/LLM/index.html).***
|
||||||
|
|
||||||
|
|
|
||||||
|
|
@ -82,6 +82,8 @@ Over 20 models have been optimized/verified on `bigdl-llm`, including *LLaMA/LLa
|
||||||
| Bark | [link](example/CPU/PyTorch-Models/Model/bark) | [link](example/GPU/PyTorch-Models/Model/bark) |
|
| Bark | [link](example/CPU/PyTorch-Models/Model/bark) | [link](example/GPU/PyTorch-Models/Model/bark) |
|
||||||
| SpeechT5 | | [link](example/GPU/PyTorch-Models/Model/speech-t5) |
|
| SpeechT5 | | [link](example/GPU/PyTorch-Models/Model/speech-t5) |
|
||||||
| Ziya-Coding-34B-v1.0 | [link](example/CPU/HF-Transformers-AutoModels/Model/ziya) | |
|
| Ziya-Coding-34B-v1.0 | [link](example/CPU/HF-Transformers-AutoModels/Model/ziya) | |
|
||||||
|
| Qwen1.5 | [link](example/CPU/HF-Transformers-AutoModels/Model/qwen1.5) | [link](example/GPU/HF-Transformers-AutoModels/Model/qwen1.5) |
|
||||||
|
|
||||||
|
|
||||||
### Working with `bigdl-llm`
|
### Working with `bigdl-llm`
|
||||||
|
|
||||||
|
|
|
||||||
|
|
@ -0,0 +1,87 @@
|
||||||
|
# Qwen1.5
|
||||||
|
|
||||||
|
In this directory, you will find examples on how you could apply BigDL-LLM INT4 optimizations on Qwen1.5 models. For illustration purposes, we utilize the [Qwen/Qwen1.5-7B-Chat](https://huggingface.co/Qwen/Qwen1.5-7B-Chat) as a reference Qwen1.5 model.
|
||||||
|
|
||||||
|
## 0. Requirements
|
||||||
|
To run these examples with BigDL-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 BigDL-LLM INT4 optimizations.
|
||||||
|
### 1. Install
|
||||||
|
We suggest using conda to manage environment:
|
||||||
|
```bash
|
||||||
|
conda create -n llm python=3.9
|
||||||
|
conda activate llm
|
||||||
|
|
||||||
|
pip install --pre --upgrade bigdl-llm[all] # install bigdl-llm with 'all' option
|
||||||
|
pip install transformers==4.37.0 # install the transformers which support Qwen2
|
||||||
|
```
|
||||||
|
|
||||||
|
### 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/Qwen1.5-7B-Chat'`.
|
||||||
|
- `--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, BigDL-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:
|
||||||
|
```powershell
|
||||||
|
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 BigDL-LLM env variables
|
||||||
|
source bigdl-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/Qwen1.5-7B-Chat](https://huggingface.co/Qwen/Qwen1.5-7B-Chat)
|
||||||
|
```log
|
||||||
|
Inference time: xxxx s
|
||||||
|
-------------------- Prompt --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
AI是什么？<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
-------------------- Output --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
AI是什么？<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
人工智能（AI）是指计算机科学的一个分支，旨在开发能够执行通常需要人类智能的任务的算法和系统。这些任务包括但不限于理解自然语言、
|
||||||
|
```
|
||||||
|
|
||||||
|
```log
|
||||||
|
Inference time: xxxx s
|
||||||
|
-------------------- Prompt --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
What is AI?<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
-------------------- Output --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
What is AI?<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
AI, or artificial intelligence, refers to the simulation of human intelligence in machines that are designed to perform tasks that typically require human cognition, such as learning, reasoning
|
||||||
|
```
|
||||||
|
|
@ -0,0 +1,77 @@
|
||||||
|
#
|
||||||
|
# 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='Qwen1.5-7B-Chat')
|
||||||
|
parser.add_argument('--repo-id-or-model-path', type=str, default="Qwen/Qwen1.5-7B-Chat",
|
||||||
|
help='The huggingface repo id for the Qwen1.5 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 bigdl.llm.transformers import AutoModelForCausalLM
|
||||||
|
model = AutoModelForCausalLM.from_pretrained(model_path,
|
||||||
|
load_in_4bit=True,
|
||||||
|
trust_remote_code=True)
|
||||||
|
|
||||||
|
# Load tokenizer
|
||||||
|
tokenizer = AutoTokenizer.from_pretrained(model_path,
|
||||||
|
trust_remote_code=True)
|
||||||
|
|
||||||
|
prompt = args.prompt
|
||||||
|
|
||||||
|
# Generate predicted tokens
|
||||||
|
with torch.inference_mode():
|
||||||
|
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 +1,86 @@
|
||||||
|
# Qwen1.5
|
||||||
|
In this directory, you will find examples on how you could use BigDL-LLM `optimize_model` API to accelerate Qwen1.5 models. For illustration purposes, we utilize the [Qwen/Qwen1.5-7B-Chat](https://huggingface.co/Qwen/Qwen1.5-7B-Chat) as reference Qwen1.5 model.
|
||||||
|
|
||||||
|
## Requirements
|
||||||
|
To run these examples with BigDL-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 Qwen1.5 model to predict the next N tokens using `generate()` API, with BigDL-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://docs.conda.io/en/latest/miniconda.html#).
|
||||||
|
|
||||||
|
After installing conda, create a Python environment for BigDL-LLM:
|
||||||
|
```bash
|
||||||
|
conda create -n llm python=3.9 # recommend to use Python 3.9
|
||||||
|
conda activate llm
|
||||||
|
|
||||||
|
pip install --pre --upgrade bigdl-llm[all] # install the latest bigdl-llm nightly build with 'all' option
|
||||||
|
pip install transformers==4.37.0 # install transformers which supports Qwen2
|
||||||
|
```
|
||||||
|
|
||||||
|
### 2. Run
|
||||||
|
After setting up the Python environment, you could run the example by following steps.
|
||||||
|
|
||||||
|
#### 2.1 Client
|
||||||
|
On client Windows machines, it is recommended to run directly with full utilization of all cores:
|
||||||
|
```powershell
|
||||||
|
python ./generate.py --prompt 'What is AI?'
|
||||||
|
```
|
||||||
|
More information about arguments can be found in [Arguments Info](#23-arguments-info) section. The expected output can be found in [Sample Output](#24-sample-output) section.
|
||||||
|
|
||||||
|
#### 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 BigDL-LLM env variables
|
||||||
|
source bigdl-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?'
|
||||||
|
```
|
||||||
|
More information about arguments can be found in [Arguments Info](#23-arguments-info) section. The expected output can be found in [Sample Output](#24-sample-output) section.
|
||||||
|
|
||||||
|
#### 2.3 Arguments Info
|
||||||
|
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 Qwen1.5 to be downloaded, or the path to the huggingface checkpoint folder. It is default to be `'Qwen/Qwen1.5-7B-Chat'`.
|
||||||
|
- `--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`.
|
||||||
|
|
||||||
|
#### 2.3 Sample Output
|
||||||
|
#### [Qwen/Qwen1.5-7B-Chat](https://huggingface.co/Qwen/Qwen1.5-7B-Chat)
|
||||||
|
```log
|
||||||
|
Inference time: xxxx s
|
||||||
|
-------------------- Prompt --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
AI是什么？<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
-------------------- Output --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
AI是什么？<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
AI（Artificial Intelligence）是指由计算机程序实现的智能，它使机器能够模拟人类的思考、学习和决策过程，从而解决各种复杂
|
||||||
|
```
|
||||||
|
|
||||||
|
```log
|
||||||
|
Inference time: xxxx s
|
||||||
|
-------------------- Prompt --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
What is AI?<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
-------------------- Output --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
What is AI?<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
AI, or artificial intelligence, refers to the simulation of human intelligence in machines that are programmed to think, learn, and make decisions like humans. It involves the
|
||||||
|
```
|
||||||
|
|
@ -0,0 +1,78 @@
|
||||||
|
#
|
||||||
|
# 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='Qwen1.5-7B-Chat')
|
||||||
|
parser.add_argument('--repo-id-or-model-path', type=str, default="Qwen/Qwen1.5-7B-Chat",
|
||||||
|
help='The huggingface repo id for the Qwen1.5 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)
|
||||||
|
|
||||||
|
|
||||||
|
# Load tokenizer
|
||||||
|
tokenizer = AutoTokenizer.from_pretrained(model_path,
|
||||||
|
trust_remote_code=True)
|
||||||
|
from bigdl.llm import optimize_model
|
||||||
|
model = optimize_model(model)
|
||||||
|
|
||||||
|
prompt = args.prompt
|
||||||
|
# Generate predicted tokens
|
||||||
|
with torch.inference_mode():
|
||||||
|
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 +1,143 @@
|
||||||
|
# Qwen1.5
|
||||||
|
In this directory, you will find examples on how you could apply BigDL-LLM INT4 optimizations on Qwen1.5 models on [Intel GPUs](../../../README.md). For illustration purposes, we utilize the [Qwen/Qwen1.5-7B-Chat](https://huggingface.co/Qwen/Qwen1.5-7B-Chat) as a reference InternLM model.
|
||||||
|
|
||||||
|
## 0. Requirements
|
||||||
|
To run these examples with BigDL-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 Qwen1.5 model to predict the next N tokens using `generate()` API, with BigDL-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=3.9
|
||||||
|
conda activate llm
|
||||||
|
# below command will install intel_extension_for_pytorch==2.1.10+xpu as default
|
||||||
|
pip install --pre --upgrade bigdl-llm[xpu] -f https://developer.intel.com/ipex-whl-stable-xpu
|
||||||
|
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=3.9 libuv
|
||||||
|
conda activate llm
|
||||||
|
# below command will install intel_extension_for_pytorch==2.1.10+xpu as default
|
||||||
|
pip install --pre --upgrade bigdl-llm[xpu] -f https://developer.intel.com/ipex-whl-stable-xpu
|
||||||
|
pip install transformers==4.37.2 # install transformers which supports Qwen2
|
||||||
|
```
|
||||||
|
|
||||||
|
### 2. Configures OneAPI environment variables
|
||||||
|
#### 2.1 Configurations for Linux
|
||||||
|
```bash
|
||||||
|
source /opt/intel/oneapi/setvars.sh
|
||||||
|
```
|
||||||
|
|
||||||
|
#### 2.2 Configurations for Windows
|
||||||
|
```cmd
|
||||||
|
call "C:\Program Files (x86)\Intel\oneAPI\setvars.bat"
|
||||||
|
```
|
||||||
|
> Note: Please make sure you are using **CMD** (**Anaconda Prompt** if using conda) to run the command as PowerShell is not supported.
|
||||||
|
### 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
|
||||||
|
```
|
||||||
|
|
||||||
|
</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 ENABLE_SDP_FUSION=1
|
||||||
|
```
|
||||||
|
> Note: Please note that `libtcmalloc.so` can be installed by `conda install -c conda-forge -y gperftools=2.10`.
|
||||||
|
</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™ A300-Series or Pro A60</summary>
|
||||||
|
|
||||||
|
```cmd
|
||||||
|
set SYCL_CACHE_PERSISTENT=1
|
||||||
|
```
|
||||||
|
|
||||||
|
</details>
|
||||||
|
|
||||||
|
<details>
|
||||||
|
|
||||||
|
<summary>For other Intel dGPU Series</summary>
|
||||||
|
|
||||||
|
There is no need to set further environment variables.
|
||||||
|
|
||||||
|
</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 Qwen1.5 model (e.g. `Qwen/Qwen1.5-7B-Chat`) to be downloaded, or the path to the huggingface checkpoint folder. It is default to be `'Qwen/Qwen1.5-7B-Chat'`.
|
||||||
|
- `--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/Qwen1.5-7B-Chat](https://huggingface.co/Qwen/Qwen1.5-7B-Chat)
|
||||||
|
```log
|
||||||
|
Inference time: xxxx s
|
||||||
|
-------------------- Prompt --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
AI是什么？<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
-------------------- Output --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
AI是什么？<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
人工智能（AI）是指通过计算机模拟、延伸和扩展人类智能的学科，其目标是使机器具有学习、推理、感知、理解、交流
|
||||||
|
```
|
||||||
|
|
||||||
|
```log
|
||||||
|
Inference time: xxxx s
|
||||||
|
-------------------- Prompt --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
What is AI?<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
-------------------- Output --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
What is AI?<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
AI, or artificial intelligence, refers to the simulation of human intelligence in machines that are programmed to perform tasks that typically require human cognition, such as learning, reasoning
|
||||||
|
```
|
||||||
|
|
@ -0,0 +1,90 @@
|
||||||
|
#
|
||||||
|
# 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 bigdl.llm import optimize_model
|
||||||
|
import intel_extension_for_pytorch as ipex
|
||||||
|
import numpy as np
|
||||||
|
|
||||||
|
|
||||||
|
if __name__ == '__main__':
|
||||||
|
parser = argparse.ArgumentParser(description='Qwen1.5-7B-Chat')
|
||||||
|
parser.add_argument('--repo-id-or-model-path', type=str, default="Qwen/Qwen1.5-7B-Chat",
|
||||||
|
help='The huggingface repo id for the Qwen1.5 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 bigdl.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,
|
||||||
|
trust_remote_code=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():
|
||||||
|
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)
|
||||||
143
python/llm/example/GPU/PyTorch-Models/Model/qwen1.5/README.md
Normal file
143
python/llm/example/GPU/PyTorch-Models/Model/qwen1.5/README.md
Normal file
|
|
@ -0,0 +1,143 @@
|
||||||
|
# Qwen1.5
|
||||||
|
In this directory, you will find examples on how you could use BigDL-LLM `optimize_model` API to accelerate Qwen1.5 models on [Intel GPUs](../../../README.md). For illustration purposes, we utilize the [Qwen/Qwen1.5-7B-Chat](https://huggingface.co/Qwen/Qwen1.5-7B-Chat) as a reference InternLM model.
|
||||||
|
|
||||||
|
## 0. Requirements
|
||||||
|
To run these examples with BigDL-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 Qwen1.5 model to predict the next N tokens using `generate()` API, with BigDL-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=3.9
|
||||||
|
conda activate llm
|
||||||
|
# below command will install intel_extension_for_pytorch==2.1.10+xpu as default
|
||||||
|
pip install --pre --upgrade bigdl-llm[xpu] -f https://developer.intel.com/ipex-whl-stable-xpu
|
||||||
|
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=3.9 libuv
|
||||||
|
conda activate llm
|
||||||
|
# below command will install intel_extension_for_pytorch==2.1.10+xpu as default
|
||||||
|
pip install --pre --upgrade bigdl-llm[xpu] -f https://developer.intel.com/ipex-whl-stable-xpu
|
||||||
|
pip install transformers==4.37.2 # install transformers which supports Qwen2
|
||||||
|
```
|
||||||
|
|
||||||
|
### 2. Configures OneAPI environment variables
|
||||||
|
#### 2.1 Configurations for Linux
|
||||||
|
```bash
|
||||||
|
source /opt/intel/oneapi/setvars.sh
|
||||||
|
```
|
||||||
|
|
||||||
|
#### 2.2 Configurations for Windows
|
||||||
|
```cmd
|
||||||
|
call "C:\Program Files (x86)\Intel\oneAPI\setvars.bat"
|
||||||
|
```
|
||||||
|
> Note: Please make sure you are using **CMD** (**Anaconda Prompt** if using conda) to run the command as PowerShell is not supported.
|
||||||
|
### 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
|
||||||
|
```
|
||||||
|
|
||||||
|
</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 ENABLE_SDP_FUSION=1
|
||||||
|
```
|
||||||
|
> Note: Please note that `libtcmalloc.so` can be installed by `conda install -c conda-forge -y gperftools=2.10`.
|
||||||
|
</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™ A300-Series or Pro A60</summary>
|
||||||
|
|
||||||
|
```cmd
|
||||||
|
set SYCL_CACHE_PERSISTENT=1
|
||||||
|
```
|
||||||
|
|
||||||
|
</details>
|
||||||
|
|
||||||
|
<details>
|
||||||
|
|
||||||
|
<summary>For other Intel dGPU Series</summary>
|
||||||
|
|
||||||
|
There is no need to set further environment variables.
|
||||||
|
|
||||||
|
</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 Qwen1.5 model (e.g. `Qwen/Qwen1.5-7B-Chat`) to be downloaded, or the path to the huggingface checkpoint folder. It is default to be `'Qwen/Qwen1.5-7B-Chat'`.
|
||||||
|
- `--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/Qwen1.5-7B-Chat](https://huggingface.co/Qwen/Qwen1.5-7B-Chat)
|
||||||
|
```log
|
||||||
|
Inference time: xxxx s
|
||||||
|
-------------------- Prompt --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
AI是什么？<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
-------------------- Output --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
AI是什么？<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
AI（Artificial Intelligence）是指计算机科学的一个分支，其目标是创建能够理解、学习、推理和自我修正的智能机器。AI系统可以通过
|
||||||
|
```
|
||||||
|
|
||||||
|
```log
|
||||||
|
Inference time: xxxx s
|
||||||
|
-------------------- Prompt --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
What is AI?<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
-------------------- Output --------------------
|
||||||
|
<|im_start|>system
|
||||||
|
You are a helpful assistant.<|im_end|>
|
||||||
|
<|im_start|>user
|
||||||
|
What is AI?<|im_end|>
|
||||||
|
<|im_start|>assistant
|
||||||
|
AI stands for Artificial Intelligence, which is a field of computer science that focuses on creating intelligent machines that can perform tasks that typically require human intelligence, such as learning
|
||||||
|
```
|
||||||
|
|
@ -0,0 +1,90 @@
|
||||||
|
#
|
||||||
|
# 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 bigdl.llm import optimize_model
|
||||||
|
import intel_extension_for_pytorch as ipex
|
||||||
|
import numpy as np
|
||||||
|
|
||||||
|
|
||||||
|
if __name__ == '__main__':
|
||||||
|
parser = argparse.ArgumentParser(description='Qwen1.5-7B-Chat')
|
||||||
|
parser.add_argument('--repo-id-or-model-path', type=str, default="Qwen/Qwen1.5-7B-Chat",
|
||||||
|
help='The huggingface repo id for the Qwen1.5 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 bigdl.llm import optimize_model
|
||||||
|
model = AutoModelForCausalLM.from_pretrained(model_path,
|
||||||
|
trust_remote_code=True,
|
||||||
|
torch_dtype = torch.float16,
|
||||||
|
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():
|
||||||
|
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