docker 服务器等操作记录

最新推荐文章于 2024-11-14 20:33:31 发布

鹿岛Eyre

最新推荐文章于 2024-11-14 20:33:31 发布

阅读量163

点赞数

文章标签： docker 服务器容器

本文链接：https://blog.csdn.net/DeerEyre/article/details/131762259

版权

conda 安装并激活到指定目录

conda create -p /mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/envs/lavis-ganshu python=3.8

conda activate /mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/envs/lavis-ganshu

conda activate /mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/envs/llama-ganshu

/mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/

ssh moxiaolin@jumper.sankuai.com

初始密码

76+@koIb)vVmYtLC

密码更改

登陆到跳板机直接输入password修改即可

jupyter notebook.

jupyter lab --config /home/hadoop-search/.jupyter/jupyter_notebook_config_ganshu.py

# hope 配置
卸载历史安装：pip uninstall hope -y

首次安装：pip install -i http://pypi.sankuai.com/simple/ hope --trusted-host pypi.sankuai.com

更新：pip install -i http://pypi.sankuai.com/simple/ hope --trusted-host pypi.sankuai.com --upgrade

streamlit run web_blip.py --server.port=8415

显存内存计算代码

from accelerate import Accelerator
import gc
import psutil
import torch
import threading
import numpy as np
def b2mb(x):
    return int(x / 2**20)
class TorchTracemalloc:
    def __enter__(self):
        gc.collect()
        torch.cuda.empty_cache()
        torch.cuda.reset_max_memory_allocated()  # reset the peak gauge to zero
        self.begin = torch.cuda.memory_allocated()
        self.process = psutil.Process()

        self.cpu_begin = self.cpu_mem_used()
        self.peak_monitoring = True
        peak_monitor_thread = threading.Thread(target=self.peak_monitor_func)
        peak_monitor_thread.daemon = True
        peak_monitor_thread.start()
        return self

    def cpu_mem_used(self):
        """get resident set size memory for the current process"""
        return self.process.memory_info().rss

    def peak_monitor_func(self):
        self.cpu_peak = -1

        while True:
            self.cpu_peak = max(self.cpu_mem_used(), self.cpu_peak)

            # can't sleep or will not catch the peak right (this comment is here on purpose)
            # time.sleep(0.001) # 1msec

            if not self.peak_monitoring:
                break

    def __exit__(self, *exc):
        self.peak_monitoring = False

        gc.collect()
        torch.cuda.empty_cache()
        self.end = torch.cuda.memory_allocated()
        self.peak = torch.cuda.max_memory_allocated()
        self.used = b2mb(self.end - self.begin)
        self.peaked = b2mb(self.peak - self.begin)

        self.cpu_end = self.cpu_mem_used()
        self.cpu_used = b2mb(self.cpu_end - self.cpu_begin)
        self.cpu_peaked = b2mb(self.cpu_peak - self.cpu_begin)
        # print(f"delta used/peak {self.used:4d}/{self.peaked:4d}")
from numpy import mean
import torch
import inspect

from torchvision import models
device = torch.device('cuda:0')
accelerator = Accelerator()
with TorchTracemalloc() as tracemalloc:
    model, vis_processors, _ = load_model_and_preprocess(name="blip_caption", model_type="base_coco", is_eval=True, device=device)
    ls =["/mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/projects/LAVIS-main/docs/_static/ganshu1.jpg"]
    memory = []
    for l in ls:
        raw_image = Image.open(l).convert("RGB")
        image = vis_processors["eval"](raw_image).unsqueeze(0).to(device)
        res=model.generate({"image": image})
        
accelerator.print("GPU Memory before entering the train : {}".format(b2mb(tracemalloc.begin)))
accelerator.print("GPU Memory consumed at the end of the train (end-begin): {}".format(tracemalloc.used))
accelerator.print("GPU Peak Memory consumed during the train (max-begin): {}".format(tracemalloc.peaked))
accelerator.print("GPU Total Peak Memory consumed during the train (max): {}".format(tracemalloc.peaked + b2mb(tracemalloc.begin)
    )
)

accelerator.print("CPU Memory before entering the train : {}".format(b2mb(tracemalloc.cpu_begin)))
accelerator.print("CPU Memory consumed at the end of the train (end-begin): {}".format(tracemalloc.cpu_used))
accelerator.print("CPU Peak Memory consumed during the train (max-begin): {}".format(tracemalloc.cpu_peaked))
accelerator.print(
    "CPU Total Peak Memory consumed during the train (max): {}".format(
        tracemalloc.cpu_peaked + b2mb(tracemalloc.cpu_begin)
    )
)


##
from numpy import mean
import torch
import inspect
import time 
from torchvision import models
device = torch.device('cuda:1')
neicun = []
xiancun = []
time_taken = []
ls =["/mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/projects/LAVIS-main/docs/_static/ganshu1.jpg","/mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/projects/LAVIS-main/docs/_static/ganshu2.jpg","/mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/projects/LAVIS-main/docs/_static/ganshu3.jpg","/mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/projects/LAVIS-main/docs/_static/ganshu4.jpg","/mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/projects/LAVIS-main/docs/_static/ganshu5.jpg","/mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/projects/LAVIS-main/docs/_static/ganshu6.jpg","/mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/projects/LAVIS-main/docs/_static/ganshu7.jpg","/mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/projects/LAVIS-main/docs/_static/ganshu8.jpg", "/mnt/dolphinfs/hdd_pool/docker/user/hadoop-search/ganshu01/projects/LAVIS-main/docs/_static/ganshu9.jpg",]
for l in ls:
    with TorchTracemalloc() as tracemalloc:
        start_time = time.time()
        raw_image = Image.open(l).convert("RGB")
        model, vis_processors, _ = load_model_and_preprocess(name="blip2_t5", model_type='caption_coco_flant5xl', is_eval=True, device=device)
        image = vis_processors["eval"](raw_image).unsqueeze(0).to(device)
        res=model.generate({"image": image})
        end_time = time.time()
    time_taken.append(end_time-start_time)
    neicun.append(tracemalloc.cpu_peaked)
    xiancun.append(tracemalloc.peaked)
print('xiancun:', xiancun)
print('neicun:', neicun)
print('time',time_taken)
print('mean_xiancun:', mean(xiancun))
print('mean_neicun:', mean(neicun))
print('mean_time:', mean(time_taken))

nohup jupyter notebook --port=8418 --config /home/hadoop-search/.jupyter/jupyter_notebook_config_ganshu.py > ganshu.log 2>&1 &

jupyter kernelspec list

jupyter kernelspec remove llama-ganshu

python -m ipykernel install --user --name llama-ganshu --display-name 'llama'

外网代理

# 添加proxy 
export http_proxy=http://10.22.139.49:6666
export https_proxy=http://10.22.139.49:6666

#验证
echo $http_proxy
#发送一条请求
curl "http://jaminzhang.github.io/nginx/Nginx-resolver-DNS-resolve-timed-out-problem-analysis-and-solve/"

netstat -anp | grep 8418