aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2023-09-18added wer to inference, transcriptsPherkel
2023-09-18Merge remote-tracking branch 'origin/improvement/readme'Pherkel
2023-09-18updated philipps configPherkel
2023-09-18small lm fixesPherkel
2023-09-18Merge pull request #38 from Algo-Boys/decoderPherkel
Decoder
2023-09-18Merge remote-tracking branch 'origin/main' into decoderPherkel
2023-09-18fixPherkel
2023-09-18added comment about changing the decoderJoJoBarthold2
2023-09-18added support for lm decoder during trainingPherkel
2023-09-18decoder changesPherkel
2023-09-18Merge branch 'decoder' of github.com:Algo-Boys/SWR2-cool-projekt into decoderPherkel
2023-09-18small fixesPherkel
2023-09-18Added swabian sampleMarvin Borner
2023-09-18Model loading fix for DataParallel modelsMarvin Borner
2023-09-18Synced cluster codeMarvin Borner
2023-09-18Added better visualizationMarvin Borner
2023-09-18reset commit historyJoJoBarthold2
2023-09-18didPherkel
2023-09-17Added swabian sampleMarvin Borner
2023-09-17Model loading fix for DataParallel modelsMarvin Borner
2023-09-16Synced cluster codeMarvin Borner
2023-09-16Added better visualizationMarvin Borner
2023-09-16created a method that prints the lexiconJoJoBarthold2
2023-09-16added a todoJoJoBarthold2
2023-09-16added a method to create a txt for the ctc decoderJoJoBarthold2
2023-09-16ignoring the downloaded modelJoJoBarthold2
2023-09-12Merge pull request #37 from Algo-Boys/fix/ultimatePherkel
Fix/ultimate
2023-09-12fixed black formatting issuePherkel
2023-09-11fix: switched up training and test splits in train.pyPhilipp Merkel
2023-09-11added cluster configPherkel
2023-09-11changed name to match pre-trained weightsPherkel
2023-09-11model loading is broken :(Pherkel
2023-09-11added n_feats from configPherkel
2023-09-11changed config to yaml!Pherkel
2023-09-11oopsie, messed up depsPherkel
2023-09-11removed bloat, reduced ci depsPherkel
2023-09-11moved own recording to data dirPherkel
2023-09-11added tokenizer to git and tokenizer training routingPherkel
2023-09-11big change!Pherkel
2023-09-11added small stuff to data utilitiesPherkel
2023-09-11refactor utilities (data, vis, tokenizer)Pherkel
2023-09-11readd space token, idk why its there but it could break thingsPherkel
2023-09-11improved tokenizerPherkel
2023-09-11improved loss_scoresPherkel
2023-09-08added siljas fileJoJoBarthold2
2023-09-06added my speechJoJoBarthold2
2023-09-06Added inferenceMarvin Borner
2023-09-06Merge pull request #34 from Algo-Boys/vaile-reads-codeJoJoBarthold2
Vaile reads code
2023-09-06fixed a spelling mistakeJoJoBarthold2
2023-09-06S makes more senseJoJoBarthold2
2023-09-06deleted a TODOJoJoBarthold2
2023-09-06added simple plotting to runJoJoBarthold2
2023-09-05Merge pull request #31 from Algo-Boys/feature/limited-supervisionPherkel
add limited supervision training (10hr)
2023-09-05add limited supervision training (10hr)Pherkel
2023-09-05Merge pull request #29 from Algo-Boys/plotMarvin
Plot
2023-09-05please add a good path for thisJoJoBarthold2
2023-09-05idk i guess this is a poetry change i didn't commitJoJoBarthold2
2023-09-05added matplotlibJoJoBarthold2
2023-09-05model now saves validation_loss and cer and werJoJoBarthold2
2023-09-04Merge pull request #27 from Algo-Boys/refactor_modularizePherkel
Refactor modularize
2023-09-04please the linterPherkel
2023-09-04applied fixes to download and tokenizersPhilipp Merkel
2023-09-03fix not unzipping datasetPherkel
2023-09-03reformat, remove depsPherkel
2023-09-03idk, hopefully this worksPherkel
2023-09-01many todosPherkel
2023-08-30wörks now!°!!Pherkel
2023-08-29added padding calculation for audiofilesPherkel
2023-08-24wip: commit before going on vacation :)Pherkel
2023-08-21adjust char tokenizer interfacePherkel
2023-08-21added simple custom dataloaderPherkel
2023-08-20Merge pull request #22 from Algo-Boys/tokenizerPherkel
Tokenizer
2023-08-20disable mypy for pipeline, as it is trashPherkel
2023-08-20pass pylint when score >9Pherkel
2023-08-20please the lintersPherkel
2023-08-20try fixing ci pipelinePherkel
2023-08-20added todosPherkel
2023-08-20Merge branch 'main' into tokenizerPherkel
2023-08-20adjusted in train.pyPherkel
2023-08-20added tokenizer trainingPherkel
2023-08-19Merge pull request #18 from Algo-Boys/not-distributed-but-still-coolMarvin
Not distributed but still cool
2023-08-19GahMarvin Borner
2023-08-19Now waiting..Marvin Borner
2023-08-19Tried some more distributingMarvin Borner
For now decided against it, so I will overwrite the changes soon.
2023-08-19Merge pull request #16 from Algo-Boys/distributorMarvin
Slurm distributor attempt one
2023-08-19Fixed some distribution thingiesMarvin Borner
2023-08-19Merge remote-tracking branch 'origin/saving' into distributorMarvin Borner
2023-08-19Started distributionMarvin Borner
2023-08-19loading now worksJoJoBarthold2
2023-08-19fixJoJoBarthold2
2023-08-19also now saves loss ( hahah funny meme)JoJoBarthold2
| || || |_
2023-08-19train now returns loss so it can be saved ( amen )JoJoBarthold2
2023-08-19saving models but not saving the lossJoJoBarthold2
2023-08-19fix DependenciesJoJoBarthold2
2023-08-19Merge branch 'wave2vec2'Pherkel
2023-08-19renamed filesPherkel
2023-08-18made linter changes (will still fail)Pherkel
2023-08-18fixed training procedurePherkel
2023-08-18fixed poetry install, added dockerfilePherkel
2023-08-18added assemblyai modelPherkel