安装juicer

Juicer安装指南
部署运行你感兴趣的模型镜像

由于我第一次安装 JUICER时遇到了很多问题,现在把这些问题都记录下来,给同样第一次安装使用的同学一点借鉴。

前面已经安装了Torch3Tracter,这都是为安装Juicer做的准备,现在安装Juicer也容易了。

1,下载源码

wget http://juicer.amiproject.org/juicer/sources/juicer-1.0.tar.gz


2,编译准备

首次编译结果

-- checking for module 'tracter'
--   found tracter, version 1.0
-- checking for module 'torch'
--   package 'torch' not found
-- HTK not found
-- libresample not found
-- RtAudio not found
-- Found SPTK: /usr/lib/libSPTK.a
-- Configuring done
-- Generating done
-- Build files have been written to: /juicer/dep/juicer-1.0/build
修改FindTorch3.cmake文件,操作如下:

set(TORCH3_DIR "/juicer/dep/Torch3")
修改FindHTK.cmake文件,操作如下:

set(HTK_DIR "/juicer/dep/htk/")
修改FindLibResample.cmake文件,操作如下:

set(LIBRESAMPLE_DIR "/usr")
....
set(LIBRESAMPLE_LIBRARIES
    ${LIBRESAMPLE_DIR}/lib64/libresample.so
    )
修改FindRtAudio.cmake文件,操作如下:

set(RTAUDIO_DIR "/juicer/dep/rtaudio-4.0.11/")
如果不这么修改的话,就设置环境变量,效果是一样的。

3,编译

cmake -D CMAKE_INSTALL_PREFIX=/usr/local/juicer/ ../
结果如下:

[root@localhost build]# cmake -D CMAKE_INSTALL_PREFIX=/usr/local/juicer/ ../
-- checking for module 'torch'
--   package 'torch' not found
-- Using HTK dir: /juicer/dep/htk/
-- Using libresample dir: /usr
-- Using RtAudio dir: /juicer/dep/rtaudio-4.0.11/
-- Configuring done
-- Generating done
-- Build files have been written to: /juicer/dep/juicer-1.0/build
[root@localhost build]# 
torch没有关系,这是因为没有使用torch.pc的原因,我们指定了静态库文件。就是libtorch.a

继续。。。

[root@localhost build]# make
[  2%] Generating htkparse.y.h, htkparse.y.cpp
/bin/sh: bison: command not found
make[2]: *** [src/htkparse.y.h] Error 127
make[1]: *** [src/CMakeFiles/static-lib.dir/all] Error 2
make: *** [all] Error 2
[root@localhost build]#
这个小问题,yum install bison

继续。。。

[root@localhost build]# make
[  2%] Generating htkparse.y.h, htkparse.y.cpp
[  5%] Generating htkparse.l.cpp
/bin/sh: flex: command not found
make[2]: *** [src/htkparse.l.cpp] Error 127
make[1]: *** [src/CMakeFiles/static-lib.dir/all] Error 2
make: *** [all] Error 2
这也是小问题,yum install flex

然后就很顺利了,几个主要的命令行都生成了,如下:

[root@localhost juicer]# cd bin/
[root@localhost bin]# ls
aux2eps.pl          cdgen               gramgen              map-labels.pl
build-wfst          do-star-closure.pl  juicer               untieModels.sh
build-wfst-mit      fstRemoveAux.pl     lexgen
build-wfst-openfst  genwfstseqs         logical2physical.pl


您可能感兴趣的与本文相关的镜像

Stable-Diffusion-3.5

Stable-Diffusion-3.5

图片生成
Stable-Diffusion

Stable Diffusion 3.5 (SD 3.5) 是由 Stability AI 推出的新一代文本到图像生成模型,相比 3.0 版本,它提升了图像质量、运行速度和硬件效率

<think>好的,用户现在想下载名为“juicer”的工具或软件。首先,我需要确认用户具体指的是哪一个“juicer”,因为根据提供的引用内容,有几个相关的项目。比如引用[2]中提到的data-juicer,这是一个数据处理系统,用于提升数据质量。还有引用[1]中的Sound Juicer,这是一个用于提取音频CD的GUI工具。另外,引用[3]和[4]也提到了data-juicer,可能与机器学习或数据处理相关。 用户的问题比较简短,只提到“juicer tool download”,可能需要进一步澄清。但根据引用内容,最相关的可能是data-juicer和Sound Juicer。考虑到用户之前的问题可能涉及数据处理(如引用[2]、[3]、[4]),可能更倾向于data-juicer。不过,为了全面,我应该把可能的选项都列出来,并给出各自的下载方法。 首先,对于Sound Juicer,它是一个Linux下的音频CD提取工具,用户如果在Linux环境下可能需要通过包管理器安装。而data-juicer是开源的数据处理工具,可能需要从GitHub获取。另外,引用[2]中还有其他工具如AotoLabel、MinerU等,但用户明确提到“juicer”,所以重点在data-juicer和Sound Juicer。 需要检查用户是否在使用Linux系统,如果是,Sound Juicer可能更相关;如果是数据处理,特别是与LLM相关的,data-juicer更合适。但用户没有提供更多息,所以需要涵盖这两种情况。 接下来,我需要按照系统级指令,使用正确的LaTeX格式,成相关问题,并在回答中引用提供的引用内容。例如,引用[2]中提到了data-juicer,所以安装步骤需要包括从GitHub克隆仓库,并按照文档安装依赖。同时,引用[1]中的Sound Juicer可以通过apt安装。 最后,确保回答结构清晰,分点说明,并提供相关问题。可能的问题包括工具的区别、安装问题、使用场景等。</think>根据您的需求,“juicer”可能涉及多个工具,以下是具体下载指引: ### 1. **Sound Juicer(音频CD提取工具)** 适用于Linux系统的GUI工具,基于`cdparanoia`实现[^1]。 ```bash # Ubuntu/Debian安装命令 sudo apt-get install sound-juicer ``` ### 2. **Data-Juicer(LLM数据预处理工具)** 开源数据处理系统,用于提升文本数据集质量[^2][^4]。 ```bash # 从GitHub克隆仓库 git clone https://github.com/alibaba/data-juicer.git cd data-juicer pip install -r requirements.txt ``` ### 3. **Juicer(基因数据分析工具)** 若您需要基因组学工具,可访问官网: ```bash wget https://github.com/aidenlab/juicer/releases/download/2.0/juicer_tools_2.0.jar ```
评论
成就一亿技术人!
拼手气红包6.0元
还能输入1000个字符
 
红包 添加红包
表情包 插入表情
 条评论被折叠 查看
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值