issues
search
hplt-project
/
OpusPocus
Marian machine translation training pipeline for thousands of models
2
stars
0
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
README and option update for the new CLI implementation
#61
varisd
closed
3 weeks ago
0
Support smarter job scheduling with Slurm runner
#60
varisd
opened
3 weeks ago
0
CLI QoL Improvements
#59
varisd
closed
3 weeks ago
0
Stopping/modifying a running pipeline should be possible without providing an explicit runner.
#58
varisd
opened
1 month ago
0
Opustrainer support + Full unit test passing (including runner tests)
#57
varisd
closed
1 month ago
0
Readme update with a full pipeline execution example
#56
varisd
closed
2 months ago
0
Init should create the pipeline directory provided by --pipeline-dir if possible
#55
varisd
opened
3 months ago
0
Fixing Slurm runner + adding runner-related unit tests
#54
varisd
closed
2 months ago
0
Reimplement TrainModelStep and GenerateVocabStep using Marian NMT Python API
#53
varisd
opened
3 months ago
0
Proper runner testing + fixing OpusPocusStep.state consistency
#52
varisd
closed
3 months ago
0
bash runner reports that "jobs were successfully submitted" but does not submit anything
#51
bhaddow
opened
3 months ago
3
Fix incorrect path for build log
#50
bhaddow
closed
3 months ago
0
Streamlining Marian NMT installation for OpusPocus
#49
varisd
closed
3 months ago
0
fixed sharding in pipeline_steps.translate + added test_translate
#48
varisd
closed
3 months ago
0
Add aggressive rules
#47
rggdmonk
closed
3 months ago
0
Add robustness towards pipeline.config change of an inited pipeline.
#46
varisd
opened
3 months ago
2
Should specify a minimum python version
#45
bhaddow
opened
3 months ago
4
Problems configuring marian
#44
bhaddow
opened
3 months ago
7
Corpus sharding simplification
#43
varisd
closed
3 months ago
0
`pip install` fails
#42
bhaddow
opened
3 months ago
1
license format
#41
bhaddow
closed
3 months ago
0
Smarter CorpusStep dataset sharding
#40
varisd
closed
3 months ago
0
CorpusStep corpus sharding requires revision
#39
varisd
closed
3 months ago
1
Unit Test Overhaul
#38
varisd
closed
4 months ago
0
Fixed DecontaminateCorpusStep (issue #33)
#37
varisd
closed
4 months ago
2
The config `pipeline.full.simple.yml` contains many errors
#36
bhaddow
closed
3 months ago
15
Improve documentation
#35
bhavitvyamalik
opened
4 months ago
6
CI/CD Testing: compiling 3rd party software on Github
#34
varisd
opened
4 months ago
0
DecontaminateStep does not properly process parallel corpora
#33
varisd
closed
3 months ago
1
Eval step implementation
#32
varisd
closed
4 months ago
0
Add proper unit tests for opuspocus.runners
#31
varisd
closed
3 months ago
1
Add COMET metric to EvaluationStep
#30
varisd
opened
4 months ago
0
Revision of OpusPocus Dependencies
#29
varisd
opened
4 months ago
0
CLI Refactor
#28
varisd
closed
4 months ago
0
Config files contain a lot of repetition - can this be avoided?
#27
bhaddow
opened
4 months ago
1
Look for Solutions for Multilingual Model Support
#26
varisd
opened
4 months ago
2
Add a Corpus Download Step
#25
varisd
opened
4 months ago
3
Hyperqueue Runner Support
#24
varisd
closed
3 weeks ago
1
Slurm Runner support
#23
varisd
closed
3 weeks ago
1
Implement PipelineConfig via Dataclasses + OmegaConf
#22
varisd
opened
4 months ago
0
Add OpusTrainer Support
#21
varisd
closed
3 weeks ago
1
Add Evaluation Step
#20
varisd
closed
4 months ago
0
CLI Refactor
#19
varisd
closed
4 months ago
0
Remove unnecessary files and configurations
#18
rggdmonk
closed
4 months ago
0
Initial Pytest + Repo Config Setup
#17
varisd
closed
5 months ago
0
Final LUMI branch merge.
#16
varisd
closed
5 months ago
0
Superbasic linting and pre-commit
#15
rggdmonk
closed
5 months ago
0
Initial Push to Main
#14
varisd
closed
5 months ago
0
`generate_vocab` keeps on running for a few HPLT datasets
#13
bhavitvyamalik
opened
8 months ago
0
Skip processing empty dataset at any stage
#12
bhavitvyamalik
opened
9 months ago
0
Next