diff --git a/ptuning/README.md b/ptuning/README.md
index f92a328..d833d1d 100644
--- a/ptuning/README.md
+++ b/ptuning/README.md
@@ -8,7 +8,7 @@
 ## 软件依赖
 运行微调需要4.27.1版本的`transformers`。除 ChatGLM-6B 的依赖之外，还需要安装以下依赖
 ```
-pip install rouge_chinese nltk jieba datasets
+pip install rouge_chinese nltk jieba datasets cpm_kernels sentencepiece accelerate
 ```
 ## 使用方法
 
diff --git a/ptuning/README_en.md b/ptuning/README_en.md
index 34a68a6..7d0b2b1 100644
--- a/ptuning/README_en.md
+++ b/ptuning/README_en.md
@@ -6,7 +6,7 @@ The following uses the [ADGEN](https://aclanthology.org/D19-1321.pdf) (advertisi
 ## Software dependencies
 Running p-tuning requires version 4.27.1 of `transformers`. In addition to the dependencies of ChatGLM-6B, the following dependencies are required
 ```
-pip install rouge_chinese nltk jieba datasets
+pip install rouge_chinese nltk jieba datasets cpm_kernels sentencepiece accelerate
 ```
 ## Instructions
 
diff --git a/ptuning/ds_train_finetune.sh b/ptuning/ds_train_finetune.sh
index 531a800..c7e6baa 100644
--- a/ptuning/ds_train_finetune.sh
+++ b/ptuning/ds_train_finetune.sh
@@ -2,8 +2,9 @@
 LR=1e-4
 
 MASTER_PORT=$(shuf -n 1 -i 10000-65535)
+NUM_GPUS=$(nvidia-smi --list-gpus | wc -l)
 
-deepspeed --num_gpus=4 --master_port $MASTER_PORT main.py \
+deepspeed --num_gpus=$NUM_GPUS --master_port $MASTER_PORT main.py \
     --deepspeed deepspeed.json \
     --do_train \
     --train_file AdvertiseGen/train.json \