74e7490fda
* Fix Baichuan2 prompt format * Fix Baichuan2 README * Change baichuan2 prompt info * Change baichuan2 prompt info
76 lines
3.3 KiB
Python
76 lines
3.3 KiB
Python
#
|
||
# 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 AutoModelForCausalLM, AutoTokenizer
|
||
from bigdl.llm import optimize_model
|
||
|
||
# prompt format referred from https://github.com/baichuan-inc/Baichuan2/issues/227
|
||
# and https://huggingface.co/baichuan-inc/Baichuan2-7B-Chat/blob/main/generation_utils.py#L7-L49
|
||
# For English prompt, you are recommended to change the prompt format.
|
||
BAICHUAN_PROMPT_FORMAT = "<reserved_106> {prompt} <reserved_107>"
|
||
|
||
if __name__ == '__main__':
|
||
parser = argparse.ArgumentParser(description='Predict Tokens using `generate()` API for Baichuan2 model')
|
||
parser.add_argument('--repo-id-or-model-path', type=str, default="baichuan-inc/Baichuan2-7B-Chat",
|
||
help='The huggingface repo id for the Baichuan2 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
|
||
|
||
# Load model
|
||
model = AutoModelForCausalLM.from_pretrained(model_path,
|
||
trust_remote_code=True,
|
||
torch_dtype='auto',
|
||
low_cpu_mem_usage=True)
|
||
|
||
# With only one line to enable BigDL-LLM optimization on model
|
||
# When running LLMs on Intel iGPUs for Windows users, we recommend setting `cpu_embedding=True` in the optimize_model function.
|
||
# This will allow the memory-intensive embedding layer to utilize the CPU instead of iGPU.
|
||
model = optimize_model(model)
|
||
|
||
model = model.to('xpu')
|
||
|
||
# Load tokenizer
|
||
tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
|
||
|
||
# Generate predicted tokens
|
||
with torch.inference_mode():
|
||
prompt = BAICHUAN2_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)
|
||
|
||
# start inference
|
||
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)
|