diff --git a/python/llm/example/GPU/HF-Transformers-AutoModels/Save-Load/.keep b/python/llm/example/GPU/HF-Transformers-AutoModels/Save-Load/.keep deleted file mode 100644 index e69de29bb2d..00000000000 diff --git a/python/llm/example/GPU/HF-Transformers-AutoModels/Save-Load/README.md b/python/llm/example/GPU/HF-Transformers-AutoModels/Save-Load/README.md new file mode 100644 index 00000000000..1fd0849943a --- /dev/null +++ b/python/llm/example/GPU/HF-Transformers-AutoModels/Save-Load/README.md @@ -0,0 +1,130 @@ +# Save/Load Low-Bit Models with BigDL-LLM Optimizations + +In this directory, you will find example on how you could save/load models with BigDL-LLM INT4 optimizations on Llama2 models on [Intel GPUs](../../../README.md). For illustration purposes, we utilize the [meta-llama/Llama-2-7b-chat-hf](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf) and [meta-llama/Llama-2-13b-chat-hf](https://huggingface.co/meta-llama/Llama-2-13b-chat-hf) as reference Llama2 models. + +## 0. Requirements +To run this example with BigDL-LLM, we have some recommended requirements for your machine, please refer to [here](../../README.md#system-support) for more information. + +## Example: Save/Load Model in Low-Bit Optimization +In the example [generate.py](./generate.py), we show a basic use case of saving/loading model in low-bit optimizations to predict the next N tokens using `generate()` API. Also, saving and loading operations are platform-independent, so you could run it on different platforms. +### 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 +``` + +#### 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 +``` + +### 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. Run +#### 3.1 Configurations for Linux +
+ +For Intel Arc™ A-Series Graphics and Intel Data Center GPU Flex Series + +```bash +export USE_XETLA=OFF +export SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=1 +``` + +
+ +
+ +For Intel Data Center GPU Max Series + +```bash +export LD_PRELOAD=${LD_PRELOAD}:${CONDA_PREFIX}/lib/libtcmalloc.so +export SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=1 +export ENABLE_SDP_FUSION=1 +``` +> Note: Please note that `libtcmalloc.so` can be installed by `conda install -c conda-forge -y gperftools=2.10`. +
+ +#### 3.2 Configurations for Windows +
+ +For Intel iGPU + +```cmd +set SYCL_CACHE_PERSISTENT=1 +set BIGDL_LLM_XMX_DISABLED=1 +``` + +
+ +
+ +For Intel Arc™ A300-Series or Pro A60 + +```cmd +set SYCL_CACHE_PERSISTENT=1 +``` + +
+ +
+ +For other Intel dGPU Series + +There is no need to set further environment variables. + +
+ +> 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 + +If you want to save the optimized low-bit model, run: +``` +python ./generate.py --save-path path/to/save/model +``` + +If you want to load the optimized low-bit model, run: +``` +python ./generate.py --load-path path/to/load/model +``` + +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 Llama2 model to be downloaded, or the path to the ModelScope checkpoint folder. It is default to be `'meta-llama/Llama-2-7b-chat-hf'`. +- `--save-path`: argument defining the path to save the low-bit model. Then you can load the low-bit directly. +- `--load-path`: argument defining the path to load low-bit model. +- `--prompt PROMPT`: argument defining the prompt to be inferred (with integrated prompt format for chat). It is default to be `'What is AI?'`. +- `--n-predict N_PREDICT`: argument defining the max number of tokens to predict. It is default to be `32`. + +#### Sample Output +#### [meta-llama/Llama-2-7b-chat-hf](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf) +```log +Inference time: xxxx s +-------------------- Output -------------------- +### HUMAN: +What is AI? + +### RESPONSE: + +AI is a term used to describe the development of computer systems that can perform tasks that typically require human intelligence, such as understanding natural language, recognizing images +``` diff --git a/python/llm/example/GPU/HF-Transformers-AutoModels/Save-Load/generate.py b/python/llm/example/GPU/HF-Transformers-AutoModels/Save-Load/generate.py new file mode 100644 index 00000000000..ee36132abe4 --- /dev/null +++ b/python/llm/example/GPU/HF-Transformers-AutoModels/Save-Load/generate.py @@ -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 +from bigdl.llm.transformers import AutoModelForCausalLM +from transformers import LlamaTokenizer + +# you could tune the prompt based on your own model, +# here the prompt tuning refers to https://huggingface.co/georgesung/llama2_7b_chat_uncensored#prompt-style +LLAMA2_PROMPT_FORMAT = """### HUMAN: +{prompt} + +### RESPONSE: +""" + +if __name__ == '__main__': + parser = argparse.ArgumentParser(description='Example of saving and loading the optimized model') + parser.add_argument('--repo-id-or-model-path', type=str, default="meta-llama/Llama-2-7b-chat-hf", + help='The huggingface repo id for the Llama2 (e.g. `meta-llama/Llama-2-7b-chat-hf` and `meta-llama/Llama-2-13b-chat-hf`) to be downloaded' + ', or the path to the huggingface checkpoint folder') + parser.add_argument('--save-path', type=str, default=None, + help='The path to save the low-bit model.') + parser.add_argument('--load-path', type=str, default=None, + help='The path to load the low-bit model.') + parser.add_argument('--prompt', type=str, default="What is AI?", + help='Prompt to infer') + parser.add_argument('--n-predict', type=int, default=32, + help='Max tokens to predict') + args = parser.parse_args() + model_path = args.repo_id_or_model_path + load_path = args.load_path + if load_path: + model = AutoModelForCausalLM.load_low_bit(load_path, trust_remote_code=True) + tokenizer = LlamaTokenizer.from_pretrained(load_path) + else: + model = AutoModelForCausalLM.from_pretrained(model_path, + load_in_4bit=True, + trust_remote_code=True) + tokenizer = LlamaTokenizer.from_pretrained(model_path, trust_remote_code=True) + + save_path = args.save_path + if save_path: + model.save_low_bit(save_path) + tokenizer.save_pretrained(save_path) + print(f"Model and tokenizer are saved to {save_path}") + + # please save/load model before you run it on GPU + model = model.to('xpu') + + # Generate predicted tokens + with torch.inference_mode(): + prompt = LLAMA2_PROMPT_FORMAT.format(prompt=args.prompt) + input_ids = tokenizer.encode(prompt, return_tensors="pt").to('xpu') + # ipex model needs a warmup, then inference time can be accurate + output = model.generate(input_ids, + max_new_tokens=args.n_predict) + + st = time.time() + output = model.generate(input_ids, + max_new_tokens=args.n_predict) + torch.xpu.synchronize() + end = time.time() + output = output.cpu() + output_str = tokenizer.decode(output[0], skip_special_tokens=True) + print(f'Inference time: {end-st} s') + print('-'*20, 'Output', '-'*20) + print(output_str)