Add GLM-4V example (#11343)
* add example * modify * modify * add line * add * add link and replace with phi-3-vision template * fix generate options * fix * fix --------- Co-authored-by: jinbridge <2635480475@qq.com>
This commit is contained in:
ivy-lv11
GitHub
parent
9b475c07db
commit
21fc781fce
6 changed files with 383 additions and 0 deletions
|
|
@ -232,6 +232,7 @@ Over 50 models have been optimized/verified on `ipex-llm`, including *LLaMA/LLaM
|
|||
| ChatGLM2 | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/chatglm2) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/chatglm2) |
|
||||
| ChatGLM3 | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/chatglm3) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/chatglm3) |
|
||||
| GLM-4 | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/glm4) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/glm4) |
|
||||
| GLM-4V | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/glm-4v) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/glm-4v) |
|
||||
| Mistral | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/mistral) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/mistral) |
|
||||
| Mixtral | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/mixtral) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/mixtral) |
|
||||
| Falcon | [link](python/llm/example/CPU/HF-Transformers-AutoModels/Model/falcon) | [link](python/llm/example/GPU/HF-Transformers-AutoModels/Model/falcon) |
|
||||
|
|
|
|||
|
|
@ -313,6 +313,13 @@ Verified Models
|
|||
<td>
|
||||
<a href="https://github.com/intel-analytics/ipex-llm/tree/main/python/llm/example/GPU/HF-Transformers-AutoModels/Model/glm4">link</a></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td>GLM-4V</td>
|
||||
<td>
|
||||
<a href="https://github.com/intel-analytics/ipex-llm/tree/main/python/llm/example/CPU/HF-Transformers-AutoModels/Model/glm-4v">link</a></td>
|
||||
<td>
|
||||
<a href="https://github.com/intel-analytics/ipex-llm/tree/main/python/llm/example/GPU/HF-Transformers-AutoModels/Model/glm-4v">link</a></td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td>Mistral</td>
|
||||
<td>
|
||||
|
|
|
|||
|
|
@ -0,0 +1,84 @@
|
|||
# GLM-4V
|
||||
|
||||
In this directory, you will find examples on how you could apply IPEX-LLM INT4 optimizations on GLM-4V models. For illustration purposes, we utilize the [THUDM/glm-4v-9b](https://huggingface.co/THUDM/glm-4v-9b) as a reference GLM-4V 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 GLM-4V 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 # recommend to use 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 torchvision tiktoken
|
||||
```
|
||||
|
||||
On Windows:
|
||||
|
||||
```cmd
|
||||
conda create -n llm python=3.11
|
||||
conda activate llm
|
||||
|
||||
pip install --pre --upgrade ipex-llm[all]
|
||||
|
||||
pip install torchvision tiktoken
|
||||
```
|
||||
|
||||
### 2. Run
|
||||
```
|
||||
python ./generate.py --repo-id-or-model-path REPO_ID_OR_MODEL_PATH --image-url-or-path IMAGE_URL_OR_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 GLM-4V model (e.g. `THUDM/glm-4v-9b`) to be downloaded, or the path to the huggingface checkpoint folder. It is default to be `'THUDM/glm-4v-9b'`.
|
||||
- `--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`.
|
||||
|
||||
> **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 GLM-4V 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
|
||||
```
|
||||
|
||||
#### 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
|
||||
#### [THUDM/glm-4v-9b](https://huggingface.co/THUDM/glm-4v-9b)
|
||||
|
||||
```log
|
||||
Inference time: xxxx s
|
||||
-------------------- Prompt --------------------
|
||||
What is in the image?
|
||||
-------------------- Output --------------------
|
||||
The image shows a young child holding up a small white teddy bear dressed in a pink
|
||||
```
|
||||
|
||||
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 +1,78 @@
|
|||
#
|
||||
# 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 os
|
||||
import time
|
||||
import torch
|
||||
import argparse
|
||||
import requests
|
||||
|
||||
from PIL import Image
|
||||
from ipex_llm.transformers import AutoModelForCausalLM
|
||||
from transformers import AutoTokenizer
|
||||
|
||||
if __name__ == '__main__':
|
||||
parser = argparse.ArgumentParser(description='Predict Tokens using `generate()` API for THUDM/glm-4v-9b model')
|
||||
parser.add_argument('--repo-id-or-model-path', type=str, default="THUDM/glm-4v-9b",
|
||||
help='The huggingface repo id for the THUDM/glm-4v-9b 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
|
||||
model = AutoModelForCausalLM.from_pretrained(model_path,
|
||||
load_in_4bit=True,
|
||||
optimize_model=True,
|
||||
trust_remote_code=True,
|
||||
use_cache=True)
|
||||
|
||||
tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
|
||||
|
||||
query = args.prompt
|
||||
if os.path.exists(image_path):
|
||||
image = Image.open(image_path)
|
||||
else:
|
||||
image = Image.open(requests.get(image_path, stream=True).raw)
|
||||
|
||||
# here the prompt tuning refers to https://huggingface.co/THUDM/glm-4v-9b/blob/main/README.md
|
||||
inputs = tokenizer.apply_chat_template([{"role": "user", "image": image, "content": query}],
|
||||
add_generation_prompt=True,
|
||||
tokenize=True,
|
||||
return_tensors="pt",
|
||||
return_dict=True) # chat mode
|
||||
inputs = inputs.to('cpu')
|
||||
|
||||
# Generate predicted tokens
|
||||
with torch.inference_mode():
|
||||
gen_kwargs = {"max_length": args.n_predict, "do_sample": True, "top_k": 1}
|
||||
st = time.time()
|
||||
outputs = model.generate(**inputs, **gen_kwargs)
|
||||
outputs = outputs[:, inputs['input_ids'].shape[1]:]
|
||||
end = time.time()
|
||||
print(f'Inference time: {end-st} s')
|
||||
output_str = tokenizer.decode(outputs[0])
|
||||
print('-'*20, 'Output', '-'*20)
|
||||
print(output_str)
|
||||
|
|
@ -0,0 +1,132 @@
|
|||
# GLM-4V
|
||||
In this directory, you will find examples on how you could apply IPEX-LLM FP8 optimizations on GLM-4V models on [Intel GPUs](../../../README.md). For illustration purposes, we utilize the [THUDM/glm-4v-9b](https://huggingface.co/THUDM/glm-4v-9b) as a reference GLM-4V 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 GLM-4V model to predict the next N tokens using `generate()` API, with IPEX-LLM FP8 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 tiktoken
|
||||
```
|
||||
|
||||
#### 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 tiktoken
|
||||
```
|
||||
|
||||
### 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 GLM-4V model (e.g. `THUDM/glm-4v-9b`) to be downloaded, or the path to the huggingface checkpoint folder. It is default to be `'THUDM/glm-4v-9b'`.
|
||||
- `--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
|
||||
#### [THUDM/glm-4v-9b](https://huggingface.co/THUDM/glm-4v-9b)
|
||||
|
||||
```log
|
||||
Inference time: xxxx s
|
||||
-------------------- Prompt --------------------
|
||||
What is in the image?
|
||||
-------------------- Output --------------------
|
||||
The image shows a young child holding up a small white teddy bear dressed in a pink
|
||||
```
|
||||
|
||||
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 +1,81 @@
|
|||
#
|
||||
# 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 os
|
||||
import time
|
||||
import torch
|
||||
import argparse
|
||||
import requests
|
||||
|
||||
from PIL import Image
|
||||
from ipex_llm.transformers import AutoModelForCausalLM
|
||||
from transformers import AutoTokenizer
|
||||
|
||||
if __name__ == '__main__':
|
||||
parser = argparse.ArgumentParser(description='Predict Tokens using `generate()` API for THUDM/glm-4v-9b model')
|
||||
parser.add_argument('--repo-id-or-model-path', type=str, default="THUDM/glm-4v-9b",
|
||||
help='The huggingface repo id for the THUDM/glm-4v-9b 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 = AutoModelForCausalLM.from_pretrained(model_path,
|
||||
load_in_4bit=True,
|
||||
optimize_model=True,
|
||||
trust_remote_code=True,
|
||||
use_cache=True).half().to('xpu')
|
||||
|
||||
tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
|
||||
|
||||
query = args.prompt
|
||||
if os.path.exists(image_path):
|
||||
image = Image.open(image_path)
|
||||
else:
|
||||
image = Image.open(requests.get(image_path, stream=True).raw)
|
||||
|
||||
# here the prompt tuning refers to https://huggingface.co/THUDM/glm-4v-9b/blob/main/README.md
|
||||
inputs = tokenizer.apply_chat_template([{"role": "user", "image": image, "content": query}],
|
||||
add_generation_prompt=True,
|
||||
tokenize=True,
|
||||
return_tensors="pt",
|
||||
return_dict=True) # chat mode
|
||||
inputs = inputs.to('xpu')
|
||||
|
||||
|
||||
# Generate predicted tokens
|
||||
with torch.inference_mode():
|
||||
gen_kwargs = {"max_length": args.n_predict, "do_sample": True, "top_k": 1}
|
||||
st = time.time()
|
||||
outputs = model.generate(**inputs, **gen_kwargs)
|
||||
outputs = outputs[:, inputs['input_ids'].shape[1]:]
|
||||
end = time.time()
|
||||
print(f'Inference time: {end-st} s')
|
||||
output_str = tokenizer.decode(outputs[0])
|
||||
print('-'*20, 'Output', '-'*20)
|
||||
print(output_str)
|
||||
Loading…
Reference in a new issue