niutrans / niutrans.nmt Goto Github PK

View Code? Open in Web Editor NEW

130.0 12.0 35.0 43.54 MB

A Fast Neural Machine Translation System developed in C++.

License: Apache License 2.0

CMake 4.69% C++ 86.39% Python 6.83% Cuda 2.09%

neural-machine-translation machine-translation transformer fast-decoding

niutrans.nmt's People

Contributors

Stargazers

Watchers

niutrans.nmt's Issues

不能分配内存

[INFO] elapsed=813.4, step=100, epoch=1, total word=713106, total batch=29584, loss=14.962, lr=1.87e-05
[ERROR] (XMem.cpp line 721): Cannot allocate the memory.
terminate called without an active exception
Aborted (core dumped)
以上为日志信息，打印第一条训练信息后就报错。您是否遇到过？

关于论文两个细节讨论

你好！
读了你们工程后面附的两篇论文，有两个问题想做一些讨论。
1、2021最优的student model表现甚至高出去年的teacher model，这块主要是做了哪些优化呢？我看到今年teacher model有增加回译，这块的收益是多少？
2、在做en-de序列蒸馏的时候是否需要将训练teacher model，使用的de-en模型输出的回译数据同样用teacher model预测出新结果后给student model学习呢？这块困惑的是如果src序列和tgt序列都来自于机器翻译，在训练的时候会不会是脏数据？

Cannot compile under cuda 11.8

It cannot pass the linking stage. Here is the full log

[main] Building folder: niutensor 
[build] Starting build
[proc] Executing command: /usr/bin/cmake --build /home/pzzzzz/MyProjects/niutensor/build --config Debug --target all --
[build] -- CUDA_TOOLKIT_ROOT: /opt/cuda
[build] -- GPU_ARCH: -arch=compute_75;-code=sm_75
[build] -- try to compile with half precision
[build] -- ARCH_FLAGS:-arch=compute_75;-code=sm_75
[build] -- CUDA_LIB_PATH:
[build] -- Generate Makefile For Executable File
[build] -- Name of Executable File: NiuTrans.NMT
[build] -- On Linux or macOS; Use CUDA
[build] -- Configuring done
[build] -- Generating done
[build] -- Build files have been written to: /home/pzzzzz/MyProjects/niutensor/build
[build] Consolidate compiler generated dependencies of target NiuTrans.NMT
[build] [  1%] Linking CXX executable /home/pzzzzz/MyProjects/niutensor/bin/NiuTrans.NMT
[build] /usr/bin/ld: /opt/cuda/lib64/libcurand_static.a(curand.o): in function `curandDestroyGenerator':
[build] curand.compute_90.cudafe1.cpp:(.text+0x9cdb): undefined reference to `culibosEnterCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0x9d18): undefined reference to `culibosLeaveCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0x9d33): undefined reference to `culibosEnterCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0x9d75): undefined reference to `culibosLeaveCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0x9d90): undefined reference to `culibosEnterCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0x9dd1): undefined reference to `culibosLeaveCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0x9e9c): undefined reference to `culibosLeaveCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0x9eaf): undefined reference to `culibosLeaveCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0x9ec2): undefined reference to `culibosLeaveCriticalSection'
[build] /usr/bin/ld: /opt/cuda/lib64/libcurand_static.a(curand.o): in function `curandCreateGenerator':
[build] curand.compute_90.cudafe1.cpp:(.text+0xc3f5): undefined reference to `culibosEnterCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0xc435): undefined reference to `culibosLeaveCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0xc464): undefined reference to `culibosLeaveCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0xc4e2): undefined reference to `culibosInitializeCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0xc570): undefined reference to `culibosEnterCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0xc5b0): undefined reference to `culibosLeaveCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0xc5e4): undefined reference to `culibosLeaveCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text+0xc6c1): undefined reference to `culibosInitializeCriticalSection'
[build] /usr/bin/ld: /opt/cuda/lib64/libcurand_static.a(curand.o): in function `curandDeviceConstants<unsigned int>::get(int)':
[build] curand.compute_90.cudafe1.cpp:(.text._ZN21curandDeviceConstantsIjE3getEi[_ZN21curandDeviceConstantsIjE3getEi]+0x1b): undefined reference to `culibosEnterCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text._ZN21curandDeviceConstantsIjE3getEi[_ZN21curandDeviceConstantsIjE3getEi]+0x3b): undefined reference to `culibosLeaveCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text._ZN21curandDeviceConstantsIjE3getEi[_ZN21curandDeviceConstantsIjE3getEi]+0xe0): undefined reference to `culibosLeaveCriticalSection'
[build] /usr/bin/ld: /opt/cuda/lib64/libcurand_static.a(curand.o): in function `curandDeviceConstants<unsigned long long>::curandDeviceConstants(void*, unsigned long, void (*)())':
[build] curand.compute_90.cudafe1.cpp:(.text._ZN21curandDeviceConstantsIyEC2EPvmPFvvE[_ZN21curandDeviceConstantsIyEC5EPvmPFvvE]+0x82): undefined reference to `culibosInitializeCriticalSection'
[build] /usr/bin/ld: /opt/cuda/lib64/libcurand_static.a(curand.o): in function `curandDeviceConstants<unsigned long long>::get(int)':
[build] curand.compute_90.cudafe1.cpp:(.text._ZN21curandDeviceConstantsIyE3getEi[_ZN21curandDeviceConstantsIyE3getEi]+0x1b): undefined reference to `culibosEnterCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text._ZN21curandDeviceConstantsIyE3getEi[_ZN21curandDeviceConstantsIyE3getEi]+0x3b): undefined reference to `culibosLeaveCriticalSection'
[build] /usr/bin/ld: curand.compute_90.cudafe1.cpp:(.text._ZN21curandDeviceConstantsIyE3getEi[_ZN21curandDeviceConstantsIyE3getEi]+0xe0): undefined reference to `culibosLeaveCriticalSection'
[build] clang-14: error: linker command failed with exit code 1 (use -v to see invocation)
[build] make[2]: *** [CMakeFiles/NiuTrans.NMT.dir/build.make:15303: /home/pzzzzz/MyProjects/niutensor/bin/NiuTrans.NMT] Error 1
[build] make[1]: *** [CMakeFiles/Makefile2:83: CMakeFiles/NiuTrans.NMT.dir/all] Error 2
[build] make: *** [Makefile:91: all] Error 2
[proc] The command: /usr/bin/cmake --build /home/pzzzzz/MyProjects/niutensor/build --config Debug --target all -- exited with code: 2 and signal: null
[build] Build finished with exit code 2
[cpptools] The build configurations generated do not contain the active build configuration. Using "" for CMAKE_BUILD_TYPE instead of "Debug" to ensure that IntelliSense configurations can be found

Add -A x64 to the cmake command not -A 64

Compile on Windows
Add -A 64 to the cmake command

This sentence may have been a clerical error.

Try this：
cmake -DUSE_CUDA=ON -DCUDA_TOOLKIT_ROOT="E:/Program Files/NVIDIA GPU Computing Toolkit/CUDA" -DGPU_ARCH=M -A x64 ..

Replace "E:/Program Files/NVIDIA GPU Computing Toolkit/CUDA" with yours own CUDA directory.

It generated a visual studio project on windows.

(XMem.cpp line 721): Cannot allocate the memory

Hi, it's an amazing project, but I cannot start work with myself data which has a bigger vocabulary size about 40000. The error is "(XMem.cpp line 721): Cannot allocate the memory", and my log is :
8 -nepoch=50
9 -maxcheckpoint=10
10 -enclayer=9
11 -declayer=1
12 -embsize=256
13 -modelsize=256
14 -nhead=8
15 -maxpos=128
16
17 encoder layer: 9
18 decoder layer: 1
19 attention heads: 8
20 model size: 256
21 source vocab size: 41056
22 target vocab size: 41056
23 [INFO] loaded 160239 sentences
24 [ERROR] (XMem.cpp line 721): Cannot allocate the memory.
25 terminate called without an active exception
26 scripts/train.deen.sh: line 28: 1006 Aborted bin/NiuTrans.NMT -dev ${deviceID} -model $modelFile -train ${dataDir}/train.bin -valid ${dataDir}/valid.bin -nepoch 50 -maxcheckpoint 10 -enclayer 9 -declayer 1 -embsize 256 -modelsize 256 -nhead 8 -maxpos 128

Could you help me solve this problem?

训练好的模型如何部署

如题，想把训练好的模型部署成WEB-API，请问有什么好的建议。

Recommend Projects

React

A declarative, efficient, and flexible JavaScript library for building user interfaces.
Vue.js

🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript

TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
TensorFlow

An Open Source Machine Learning Framework for Everyone
Django

The Web framework for perfectionists with deadlines.
Laravel

A PHP framework for web artisans
D3

Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

javascript

JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
web

Some thing interesting about web. New door for the world.
server

A server is a program made to process requests and deliver data to clients.
Machine learning

Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Visualization

Some thing interesting about visualization, use data art
Game

Some thing interesting about game, make everyone happy.

Recommend Org

Facebook

We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft

Open source projects and samples from Microsoft.
Google

Google ❤️ Open Source for everyone.
Alibaba

Alibaba Open Source for everyone
D3

Data-Driven Documents codes.
Tencent

China tencent open source team.