47e90a362f
* Add --modelscope for more models * imporve readme --------- Co-authored-by: ATMxsp01 <shou.xu@intel.com>
89 lines
3.7 KiB
Python
89 lines
3.7 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
|
||
import numpy as np
|
||
|
||
from ipex_llm.transformers import AutoModel
|
||
|
||
# you could tune the prompt based on your own model,
|
||
# here the prompt tuning refers to https://huggingface.co/THUDM/glm-4-9b-chat/blob/main/tokenization_chatglm.py
|
||
GLM4_PROMPT_FORMAT = "<|user|>\n{prompt}\n<|assistant|>"
|
||
|
||
if __name__ == '__main__':
|
||
parser = argparse.ArgumentParser(description='Predict Tokens using `generate()` API for GLM-4 model')
|
||
parser.add_argument('--repo-id-or-model-path', type=str,
|
||
help='The Hugging Face or ModelScope repo id for GLM-4 model model to be downloaded'
|
||
', or the path to the 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')
|
||
parser.add_argument('--modelscope', action="store_true", default=False,
|
||
help="Use models from modelscope")
|
||
|
||
args = parser.parse_args()
|
||
|
||
if args.modelscope:
|
||
from modelscope import AutoTokenizer
|
||
model_hub = 'modelscope'
|
||
else:
|
||
from transformers import AutoTokenizer
|
||
model_hub = 'huggingface'
|
||
|
||
model_path = args.repo_id_or_model_path if args.repo_id_or_model_path else \
|
||
("ZhipuAI/glm-4-9b-chat" if args.modelscope else "THUDM/glm-4-9b-chat")
|
||
|
||
# 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 = AutoModel.from_pretrained(model_path,
|
||
load_in_4bit=True,
|
||
optimize_model=True,
|
||
trust_remote_code=True,
|
||
use_cache=True,
|
||
model_hub=model_hub)
|
||
model = model.half().to("xpu")
|
||
|
||
# Load tokenizer
|
||
tokenizer = AutoTokenizer.from_pretrained(model_path,
|
||
trust_remote_code=True)
|
||
|
||
# Generate predicted tokens
|
||
with torch.inference_mode():
|
||
prompt = GLM4_PROMPT_FORMAT.format(prompt=args.prompt)
|
||
input_ids = tokenizer.encode(prompt, return_tensors="pt").to('xpu')
|
||
|
||
# ipex_llm 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_str = tokenizer.decode(output[0], skip_special_tokens=True)
|
||
print(f'Inference time: {end-st} s')
|
||
print('-'*20, 'Prompt', '-'*20)
|
||
print(prompt)
|
||
print('-'*20, 'Output', '-'*20)
|
||
print(output_str)
|