run_model get error when benchmark native int4 on CPU

[violet17]

Traceback (most recent call last):
  File "/home/a770/crystal/benchmark/all-in-one/run.py", line 544, in <module>
    run_model(model, api, conf['in_out_pairs'], conf['local_model_hub'], conf['warm_up'], conf['num_trials'], conf['num_beams'])
  File "/home/a770/crystal/benchmark/all-in-one/run.py", line 60, in run_model
    np.mean(result[in_out_pair], axis=0)[0],
UnboundLocalError: local variable 'result' referenced before assignment

benchmark native int4 on CPU:
python run.py

bigdl-llm version: 2.4.0b20231015

[shane-huang]

@qiuxin2012 please kindly take a look.

[qiuxin2012]

PR #9163 can skip this error.
We need to improve the output of native int api by #9133 .

[qiuxin2012]

#9163 is merged, you can update your run.py to upstream, and try again.

[violet17]

Thanks. @qiuxin2012 which version should I use?

[qiuxin2012]

Thanks. @qiuxin2012 which version should I use?

The latest https://github.com/intel-analytics/BigDL/blob/main/python/llm/dev/benchmark/all-in-one/run.py