NVIDIA/OpenSeq2Seq issues and pull requests

#100 - Docs fixes

Pull Request - State: closed - Opened by Kipok over 6 years ago

#99 - Fix using existing models docs

Pull Request - State: closed - Opened by Kipok over 6 years ago

#98 - Move learning rate to lr_poilcy_params

Pull Request - State: closed - Opened by Kipok over 6 years ago

#97 - Docs changes

Pull Request - State: closed - Opened by Kipok over 6 years ago

#96 - Renaming of base classes

Pull Request - State: closed - Opened by Kipok over 6 years ago

#95 - Add regularize_bn parameter

Pull Request - State: closed - Opened by Kipok over 6 years ago

#94 - Test configs

Pull Request - State: closed - Opened by Kipok over 6 years ago

#93 - Move resnet parameters to config, enable float16 computation on Horovod

Pull Request - State: closed - Opened by Kipok over 6 years ago

#92 - API changes: input_tensors is dict instead of list, also added ResNet-50 model

Pull Request - State: closed - Opened by Kipok over 6 years ago

#91 - Documentation fixes and new (old) LARC implementation

Pull Request - State: closed - Opened by Kipok over 6 years ago

#90 - Fix unittests for speech2text

Pull Request - State: closed - Opened by Kipok over 6 years ago

#89 - Fix data_size bug

Pull Request - State: closed - Opened by Kipok over 6 years ago

#88 - Dev0.3

Pull Request - State: closed - Opened by Kipok over 6 years ago

#87 - WIP: New data_layer API

Pull Request - State: closed - Opened by Kipok over 6 years ago

#86 - Transformer work

Pull Request - State: closed - Opened by okuchaiev over 6 years ago

#85 - API changes to support evaluation on horovod

Pull Request - State: closed - Opened by Kipok over 6 years ago

#84 - Updated docs (getting started -> asr)

Pull Request - State: closed - Opened by vsl9 over 6 years ago

#83 - Add start_experiment.sh functionality to run.py

Pull Request - State: closed - Opened by Kipok over 6 years ago - 4 comments

#82 - Fix epoch number for horovod

Pull Request - State: closed - Opened by Kipok over 6 years ago

#81 - Add epoch number and fixed lr summaries

Pull Request - State: closed - Opened by Kipok over 6 years ago

#80 - Changes to LARC algorightm

Pull Request - State: closed - Opened by Kipok over 6 years ago

#79 - Fix renaming bug

Pull Request - State: closed - Opened by Kipok over 6 years ago

#78 - Add resnet model, add bug fixes from master

Pull Request - State: closed - Opened by Kipok over 6 years ago - 9 comments

#77 - Fix unicode error in python 2

Pull Request - State: closed - Opened by Kipok over 6 years ago

#76 - fix in attention: bias after sqrt_normalize

Pull Request - State: closed - Opened by borisgin over 6 years ago

#75 - Add Resnet50-based encoder to OpenSeq2Seq

Issue - State: closed - Opened by okuchaiev over 6 years ago - 4 comments
Labels: new-feature

#74 - modified normaliztion to support batch_norm for fp16

Pull Request - State: closed - Opened by borisgin over 6 years ago

#73 - Fix to Issue 62

Pull Request - State: closed - Opened by okuchaiev over 6 years ago

#72 - Fix python2 compatibility

Pull Request - State: closed - Opened by Kipok over 6 years ago

#71 - Evaluation numbers are not exact when using num_gpus > 1

Issue - State: closed - Opened by Kipok over 6 years ago - 2 comments

#70 - Fix all deprecation warning

Issue - State: closed - Opened by okuchaiev over 6 years ago - 2 comments
Labels: enhancement, good first issue

#69 - DeepSpeech2 encoder treats padding as valid input

Issue - State: closed - Opened by Kipok over 6 years ago
Labels: enhancement

#68 - Allow RNN decoders to tie embedding and projection weights

Issue - State: closed - Opened by okuchaiev over 6 years ago
Labels: enhancement, good first issue

#67 - R0.2 18.05

Pull Request - State: closed - Opened by Kipok over 6 years ago

#66 - R0.2 18.05

Pull Request - State: closed - Opened by Kipok over 6 years ago

#65 - Docs changes

Pull Request - State: closed - Opened by Kipok over 6 years ago

#64 - Refactor old functions

Pull Request - State: closed - Opened by Kipok over 6 years ago

#63 - Updated LibriSpeech dataset downloader

Pull Request - State: closed - Opened by vsl9 over 6 years ago

#62 - All text2text embedding and projections should be able to (optionally) pad to 8

Issue - State: closed - Opened by okuchaiev over 6 years ago - 1 comment
Labels: enhancement

#61 - Make RNN encoders and decoders accept arbitrary TF RNN cell from config

Issue - State: closed - Opened by okuchaiev over 6 years ago - 1 comment
Labels: enhancement

#60 - Remove alphabet class

Pull Request - State: closed - Opened by Kipok over 6 years ago - 1 comment

#59 - Docs changes

Pull Request - State: closed - Opened by Kipok over 6 years ago

#58 - Docs changes

Pull Request - State: closed - Opened by Kipok over 6 years ago

#57 - docs rebuild

Pull Request - State: closed - Opened by Kipok over 6 years ago

#56 - Bugfix for beam_search in mixed precision/float16 modes

Pull Request - State: closed - Opened by okuchaiev over 6 years ago

#55 - Bugfix for MP beam search

Pull Request - State: closed - Opened by okuchaiev over 6 years ago

#54 - Merge master to 18.05

Pull Request - State: closed - Opened by okuchaiev over 6 years ago

#53 - Bugfix for beam_search in mixed precision/float16 modes

Pull Request - State: closed - Opened by okuchaiev over 6 years ago - 1 comment

#52 - Beam search decoder does not work in mixed precision/float16

Issue - State: closed - Opened by okuchaiev over 6 years ago - 1 comment
Labels: bug

#51 - Transformer bn

Pull Request - State: closed - Opened by borisgin over 6 years ago

#50 - Added BatchNorm to Transformer;

Pull Request - State: closed - Opened by borisgin over 6 years ago - 1 comment

#49 - Api change

Pull Request - State: closed - Opened by Kipok over 6 years ago - 1 comment

#48 - Added DS2 models

Pull Request - State: closed - Opened by vsl9 over 6 years ago

#47 - API changes

Pull Request - State: closed - Opened by Kipok over 6 years ago - 2 comments

#46 - Add Prerequisites to mixed precision doc section

Pull Request - State: closed - Opened by okuchaiev over 6 years ago

#45 - Docs for model, encoder, decoder classes

Pull Request - State: closed - Opened by Kipok over 6 years ago

#44 - Dockerfile: changed path to gitlab-dl

Pull Request - State: closed - Opened by borisgin over 6 years ago

#43 - Docs changes

Pull Request - State: closed - Opened by Kipok over 6 years ago

#42 - R0.2

Pull Request - State: closed - Opened by okuchaiev over 6 years ago

#41 - Change name font, fix typos

Pull Request - State: closed - Opened by Kipok over 6 years ago

#40 - Docs changes

Pull Request - State: closed - Opened by Kipok over 6 years ago

#39 - Add logo to docs

Pull Request - State: closed - Opened by Kipok over 6 years ago

#38 - Rename config parameters

Pull Request - State: closed - Opened by Kipok over 6 years ago

#37 - change index page a little

Pull Request - State: closed - Opened by okuchaiev over 6 years ago

#36 - Add model pointer to encoder, decoder and loss classes

Issue - State: closed - Opened by Kipok over 6 years ago - 1 comment
Labels: enhancement

#35 - R0.2 docs mp

Pull Request - State: closed - Opened by borisgin over 6 years ago

#34 - R0.2 docs

Pull Request - State: closed - Opened by borisgin over 6 years ago

#33 - R0.2 docs

Pull Request - State: closed - Opened by borisgin over 6 years ago

#32 - Change docs structure, add debug parameter

Pull Request - State: closed - Opened by Kipok over 6 years ago

#31 - Fill out models and recipies section in docs

Issue - State: closed - Opened by Kipok over 6 years ago - 1 comment
Labels: docs

#30 - Fill out mixed precision section in docs

Issue - State: closed - Opened by Kipok over 6 years ago - 1 comment

#29 - Add contributing.md and authors.md

Issue - State: closed - Opened by Kipok over 6 years ago - 1 comment

#28 - Mixed precision docs, new random speech data layer

Pull Request - State: closed - Opened by Kipok over 6 years ago

#27 - R0.2 transformer work

Pull Request - State: closed - Opened by okuchaiev over 6 years ago

#26 - Change docs

Pull Request - State: closed - Opened by Kipok over 6 years ago

#25 - New docs structure

Pull Request - State: closed - Opened by Kipok over 6 years ago

#24 - Add some documentation, fix benchmarking, add start_experiment.sh

Pull Request - State: closed - Opened by Kipok over 6 years ago

#23 - model.py documentation, test fixes

Pull Request - State: closed - Opened by Kipok over 6 years ago

#22 - Restructure API documentation, rename training->running

Pull Request - State: closed - Opened by Kipok over 6 years ago

#21 - Adjust lr policy for transformers

Pull Request - State: closed - Opened by okuchaiev over 6 years ago

#20 - Fix bug with + 1 in loop, tests work on 1 gpu, new params for bn

Pull Request - State: closed - Opened by Kipok over 6 years ago

#19 - update docs2

Pull Request - State: closed - Opened by okuchaiev over 6 years ago

#18 - update docs

Pull Request - State: closed - Opened by okuchaiev over 6 years ago

#17 - remove unnecessary projection layers from skip connections when they …

Pull Request - State: closed - Opened by okuchaiev almost 7 years ago

#16 - Mt session

Pull Request - State: closed - Opened by okuchaiev almost 7 years ago

#15 - addin support for distributed training using Horovod

Pull Request - State: closed - Opened by okuchaiev almost 7 years ago

#14 - Added a script to allows new users to run the GNMT like model on wmt1…

Pull Request - State: closed - Opened by matanhs almost 7 years ago - 5 comments

#13 - toy data now in unicode.

Pull Request - State: closed - Opened by nluehr about 7 years ago

#12 - More Python 2.7 fixes for german > english example

Pull Request - State: closed - Opened by nluehr about 7 years ago

#11 - Add Python 2.7 support to open_seq2seq.

Pull Request - State: closed - Opened by nluehr about 7 years ago - 1 comment

#10 - add slstm cell support

Pull Request - State: closed - Opened by okuchaiev about 7 years ago

#9 - Gnmt like

Pull Request - State: closed - Opened by okuchaiev about 7 years ago

#8 - Release new pre-trained models

Issue - State: closed - Opened by okuchaiev about 7 years ago - 1 comment

#7 - Add support for GNMT-like model

Issue - State: closed - Opened by okuchaiev about 7 years ago - 2 comments

#6 - Adding LARS implementation.

Pull Request - State: closed - Opened by okuchaiev about 7 years ago

#5 - multi-gpu example and pre-trained model

Pull Request - State: closed - Opened by okuchaiev about 7 years ago

#4 - add QA

Pull Request - State: closed - Opened by LouisChen1992 about 7 years ago

#3 - Added stdout default for inference

Pull Request - State: closed - Opened by madrugado about 7 years ago - 2 comments

#2 - Pretrained Model

Issue - State: closed - Opened by syed-ahmed about 7 years ago - 2 comments

#1 - Updating README and adding references

Pull Request - State: closed - Opened by siddharthbhatnagar about 7 years ago

GitHub / NVIDIA/OpenSeq2Seq issues and pull requests