Comments (3)
Kinda figured it out going through the code base, I think it runs 5 iterations by default. Takes about 20 min /iteration on a t2.xlarge machine.
Mohammed Ayub
from fast_align.
I have problem with larger dataset while using ./fast_align
I found that I could not deal with the data more than 1000 lines maybe. Have you ever met the same problem or how do you usually run the script?
For me it is:
./fast_align -i ~/sockeye_autopilot/systems/wmt14_en_de/data/bpe/new_combined_srctrgaa -d -o -v > forward.align_endea
Thanks a lot!
from fast_align.
@Ryanight I have no problem running files with more than 1000 lines. Maybe the input file has some problems. I have created the input file and the running the commands like below:
create the input file:
pr -mtJS' ||| ' /home/ubuntu/mayub/datasets/train_tokenized_bpe_applied.en /home/ubuntu/mayub/datasets/train_tokenized_bpe_applied.es > /home/ubuntu/mayub/datasets/fastalign_input_en_es.txt
forward train:
fast_align/build/fast_align -i /home/ubuntu/mayub/datasets/fastalign_input_en_es.txt -d -o -v > /home/ubuntu/mayub/datasets/forward_align_en_es.align
reverse train:
fast_align/build/fast_align -i /home/ubuntu/mayub/datasets/fastalign_input_en_es.txt -d -o -v -r > /home/ubuntu/mayub/datasets/reverse_align_en_es.align
symmetrize:
fast_align/build/atools -i /home/ubuntu/mayub/datasets/forward_align_en_es.align -j /home/ubuntu/mayub/datasets/reverse_align_en_es.align -c grow-diag-final-and > /home/ubuntu/mayub/datasets/corpus_en_es.gdfa
Unfortunately i have not found a good tool that can detect errors or troubleshoot the above process if the files contain error in alignments, which may in some cases. Open to suggestions.
from fast_align.
Related Issues (20)
- [BUG] tot_len_ratio being inf when src.size() is zero
- Problem with large data HOT 5
- force_alignment with score HOT 1
- How can I reproduce the result reported in the paper ?
- Inference on new corpus by trained alignments HOT 7
- en-jap alignements HOT 2
- Issue when symmetrizing forward and reversed links HOT 1
- How to train this model in a new corpus? HOT 5
- Error happened: expected target length = source length * 0.934537 By train a large dataset HOT 1
- Setting up fast_align HOT 2
- Is it possible to hide stdout when aligning?
- Not able to train model for chinese HOT 1
- Implementing fast_align model in detail!!
- some questions
- force_align.py (TypeError: a bytes-like object is required, not 'str') HOT 3
- bash: ./fast_align: 没有那个文件或目录
- Alignments generated by force_align is different from the original alignment. HOT 2
- some confusions
- What does this Error message mean ? HOT 1
- Paper link in README is expired HOT 1
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.
from fast_align.