去掉多余依赖包

Dockerfile

+FROM image.sourcefind.cn:5000/dcu/admin/base/pytorch:1.13.1-centos7.6-dtk-23.04-py39-latest
+COPY requirements.txt requirements.txt
+RUN source /opt/dtk-23.04/env.sh
+ENV LANG C.UTF-8
+RUN pip install -r requirements.txt -i http://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com
+

README.md

@@ -30,20 +30,46 @@ https://arxiv.org/pdf/2308.12966.pdf
 
 ## 环境配置
 
+### Docker(方式一)
+
 推荐使用docker方式运行，提供[光源](https://www.sourcefind.cn/#/main-page)拉取的docker镜像：
 
 ```
 docker pull image.sourcefind.cn:5000/dcu/admin/base/pytorch:1.13.1-centos7.6-dtk-23.04-py39-latest
 
 docker run -dit --network=host --name=qwen_pytorch --privileged --device=/dev/kfd --device=/dev/dri --ipc=host --shm-size=16G  --group-add video --cap-add=SYS_PTRACE --security-opt seccomp=unconfined -u root --ulimit stack=-1:-1 --ulimit memlock=-1:-1  image.sourcefind.cn:5000/dcu/admin/base/pytorch:1.13.1-centos7.6-dtk-23.04-py39-latest
+docker exec -it qwen_pytorch /bin/bash
+pip install -r requirements.txt  -i https://mirrors.aliyun.com/pypi/simple/  --trusted-host mirrors.aliyun.com
+```
+
+## Dockerfile(方式二)
+
+```
+docker build -t qwen:latest .
+docker run -dit --network=host --name=qwen_pytorch --privileged --device=/dev/kfd --device=/dev/dri --ipc=host --shm-size=16G  --group-add video --cap-add=SYS_PTRACE --security-opt seccomp=unconfined -u root --ulimit stack=-1:-1 --ulimit memlock=-1:-1 qwen:latest
+docker exec -it qwen_pytorch /bin/bash
+pip install -r requirements.txt  -i https://mirrors.aliyun.com/pypi/simple/  --trusted-host mirrors.aliyun.com
 ```
 
-进入docker
+### conda(方式三)
 
 ```
+conda create -n qwen python=3.9
 pip install -r requirements.txt  -i https://mirrors.aliyun.com/pypi/simple/  --trusted-host mirrors.aliyun.com
+```
+
+[torch1.13-dtk23.04](https://cancon.hpccube.com:65024/directlink/4/pytorch/dtk23.04/torch-1.13.1+git55d300e.abi0.dtk2304-cp39-cp39-manylinux2014_x86_64.whl)
+
+[deepspeed0.9.2-dtk23.04](https://cancon.hpccube.com:65024/directlink/4/deepspeed/dtk23.04/deepspeed-0.9.2+git25d5540.abi0.dtk2304.torch1.13.1-cp39-cp39-manylinux2014_x86_64.whl)
+
+Tips：以上dtk驱动、python、deepspeed等工具版本需要严格一一对应。
 
-#注释掉一些版本判断
+### 注意
+
+由于dtk版本的deepspeed目前最高是0.9.2因此需要进入虚拟环境修改一些版本判断
+
+```
+#到虚拟环境下对应的python/site-packages注释掉一些版本判断
 site-packages/accelerate/accelerator.py 文件
 
  287             #if not is_deepspeed_available():
@@ -51,7 +77,7 @@ site-packages/accelerate/accelerator.py 文件
  289             #if compare_versions("deepspeed", "<", "0.9.3"):
  290             #    raise ImportError("DeepSpeed version must be >= 0.9.3. Please update DeepSpeed.")
  
- site-packages/transformers/utils/versions.py 文件
+site-packages/transformers/utils/versions.py 文件
  43     #if not ops[op](version.parse(got_ver), version.parse(want_ver)):
  44     #    raise ImportError(
  45     #        f"{requirement} is required for a normal functioning of this module, but found {pkg}=={got_ver}.{hint}"
@@ -110,11 +136,11 @@ sh mpirun-nodes.sh
 
 ### 算法类别
 
-`自然语言处理`
+`对话问答`
 
 ### 热点应用行业
 
-`nlp，智能聊天助手`
+`科研`
 
 ## 源码仓库及问题反馈
 

deepspeed.json

 {
-    "train_micro_batch_size_per_gpu": "auto",
-    "zero_allow_untested_optimizer": true,
-    "fp16": {
-      "enabled": true,
-      "loss_scale": 0,
-      "initial_scale_power": 11,
-      "loss_scale_window": 1000,
-      "hysteresis": 2,
-      "min_loss_scale": 1
+  "train_micro_batch_size_per_gpu": "auto",
+  "zero_allow_untested_optimizer": true,
+  "fp16": {
+    "enabled": true,
+    "loss_scale": 0,
+    "initial_scale_power": 11,
+    "loss_scale_window": 1000,
+    "hysteresis": 2,
+    "min_loss_scale": 1
+  },
+  "zero_optimization": {
+    "stage": 3,
+    "allgather_partitions": true,
+    "allgather_bucket_size": 5e8,
+    "overlap_comm": false,
+    "reduce_scatter": true,
+    "reduce_bucket_size": 5e8,
+    "contiguous_gradients" : true,
+    "offload_optimizer": {
+      "device": "cpu",
+      "pin_memory": true
     },
-    "zero_optimization": {
-      "stage": 3,
-      "allgather_partitions": true,
-      "allgather_bucket_size": 5e8,
-      "overlap_comm": false,
-      "reduce_scatter": true,
-      "reduce_bucket_size": 5e8,
-      "contiguous_gradients" : true
-    },
-   "gradient_accumulation_steps": "auto",
-      "gradient_clipping": "auto",
-      "steps_per_print": 2000,
-      "train_batch_size":  "auto",
-      "train_micro_batch_size_per_gpu": 1,
-      "wall_clock_breakdown": false
-  }
-  
+    "offload_param": {
+      "device": "cpu",
+      "pin_memory": true
+    }
+  },
+ "gradient_accumulation_steps": "auto",
+    "gradient_clipping": "auto",
+    "steps_per_print": 2000,
+    "train_batch_size":  "auto",
+    "train_micro_batch_size_per_gpu": 1,
+    "wall_clock_breakdown": false
+}
+

model.properties

@@ -5,6 +5,6 @@ modelName=qwen_pytorch
 # 模型描述
 modelDescription=Qwen是阿里开源的预训练语言表征模型。
 # 应用场景
-appScenario=训练,NLP,文本问答
+appScenario=训练,NLP,对话问答
 # 框架类型
 frameType=Pytorch

mpirun-nodes.sh

@@ -6,6 +6,6 @@ np=$(($np*8))
 nodename=$(cat $hostfile |sed -n "1p")
 dist_url=`echo $nodename | awk '{print $1}'`
 which mpirun
-mpirun -np $np --allow-run-as-root --hostfile $hostfile --bind-to none --mca btl_tcp_if_include $dist_url single-16B.sh $dist_url
+mpirun -np $np --allow-run-as-root --hostfile $hostfile --bind-to none --mca btl_tcp_if_include $dist_url single_ddp.sh $dist_url
 echo "END TIME: $(date)"
 

requirements.txt

@@ -12,7 +12,6 @@ click==8.1.7
 contourpy==1.1.1
 cycler==0.12.1
 datasets==2.14.5
-deepspeed @ file:///work/home/hepj/DTK-whl/dtk23.04/deepspeed-0.9.2%2Bgit25d5540.abi0.dtk2304.torch1.13.1-cp39-cp39-manylinux2014_x86_64.whl
 dill==0.3.7
 docstring-parser==0.15
 einops==0.6.1
@@ -83,7 +82,6 @@ starlette==0.26.1
 tiktoken==0.5.1
 tokenizers==0.13.3
 toolz==0.12.0
-torch @ file:///work/home/hepj/DTK-whl/dtk23.04/torch-1.13.1%2Bgit55d300e.abi0.dtk2304-cp39-cp39-manylinux2014_x86_64.whl
 tqdm==4.66.1
 transformers==4.31.0
 transformers-stream-generator==0.0.4