Releases: pythonlessons/mltu
Releases · pythonlessons/mltu
Release list
1.1.4
[1.1.4] - 2022-09-29
Changed
- Improoved
mltu.torch.dataProvider.DataProviderto hanglemultiprocessingwhen it doesn't work to switch tomultithreading
1.1.3
[1.1.3] - 2022-09-29
Changed
- Removed
Librosalibrary dependency in requirements, now it is optional and required only with modules that use librosa
Added
- Created
Tutorials.05_sound_to_text.train_no_limit.pythat demonstrates how to train audio recognition model withmltuwithout audio length limit
1.1.1
[1.1.1] - 2022-09-26
Changed
- Included
self._executoras generator inmltu.dataProvider.DataProviderobject, to enable functionality to modify batch preprocessing without changing original code - Introduced changes in
mltu.torch.dataProvider.pyto handle data in multiprocessing and multithreading modes, for faster preprocessing while torch models - Modified
mltu.transformers.AudioPaddingobject, to work with batches of raw audio data
Added
- Created tutorial
10_wav2vec2_torch(Audio to Text model) that shows how to train wav2vec2 model with mltu
1.1.0
[1.1.0] - 2022-08-28
Changed
- Changed
mltu.transformers.SpectrogramPaddingobject, to pad spectrogram end with zeros instead of start
Added
- Created
Tutorials/09_translation_transformertutorial, that shows how to train translation transformer model - Created
mltu.tensorflow.tokenizersmodule, that containsCustomTokenizerfor text data - Created
mltu.tensorflow.transformer.attentionmodule, that containsBaseAttention,CrossAttention,GlobalSelfAttentionandCausalSelfAttentionlayers - Created
mltu.tensorflow.transformer.layersmodule, that containspositional_encodingfunction,PositionalEmbedding,FeedForward,EncoderLayer,DecoderLayer,Encoder,Decoderlayers andTransformermodel - Created
mltu.tensorflow.transformer.callbacksmodule, that containsEncDecSplitCallbackcallback, to split Transformer model into separate encoder and decoder models - Created
mltu.tensorflow.transformer.utilsmodule, that containsMaskedLossloss andMaskedAccuracymetric, used for training Transformer models
1.0.15
[1.0.15] - 2022-07-15
Changed
- Fixed bug in
mltu.dataProvider.DataProviderto work withbatch_postprocessors.
1.0.14
[1.0.14] - 2022-07-13
Changed
- Included
augment_annotationbool option to allmltu.augmentorsto be able to choose whether to augment annotation or not - Changed
mltu.augmentors.RandomRotateto have@staticmethodofrotate_imageto be able to use it without creating object
Added
- Added
batch_postprocessoroption tomltu.dataProvider.DataProviderto be able to postprocess batch after augmentation
1.0.12
1.0.11
1.0.10
1.0.9
Spelling mistake fixes, single quotes to double quotes, introduced CVImage and PillowImage objects