autotune api update (#2850)

Author: ZailiWang

examples/cpu/inference/python/llm/llm_sq_recipes.md

@@ -12,7 +12,6 @@ OMP_NUM_THREADS=56 numactl -m 0 -C 0-55 python run.py  --benchmark -m meta-llama
 ## Example command for model tuning with AutoTune API
 | Model ID | Command |
 |---|:---:|
-| meta-llama/Llama-2-7b-hf | python run.py -m meta-llama/Llama-2-7b-hf --ipex-smooth-quant --batch-size 56 --calib-len 2048 --fallback-add --alpha auto --init-alpha 0.8 --alpha-min 0.8 --alpha-max 0.99 --alpha-step 0.01 --shared-criterion 'mean' |
 | meta-llama/Llama-2-13b-hf | python run.py -m meta-llama/Llama-2-13b-hf --ipex-smooth-quant --alpha auto --init-alpha 0.8 --alpha-min 0.75 --alpha-max 0.99 --alpha-step 0.01 --shared-criterion 'max' --calib-len 1024 --calib-padding --fallback-add |
 | meta-llama/Llama-2-70b-hf | python run.py -m meta-llama/Llama-2-70b-hf --ipex-smooth-quant --batch-size 56 --calib-shuffle --fallback-add --alpha 0.8 |
 | EleutherAI/gpt-j-6b | python run.py -m EleutherAI/gpt-j-6b --ipex-smooth-quant --batch-size 56 --calib-iters 100 --calib-shuffle --fallback-add --alpha 0.85 |