Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / sdrobert/pydrobert-pytorch issues and pull requests
#82 - Doc fixes
Pull Request -
State: closed - Opened by sdrobert 10 months ago
#81 - Compatibility fixes for param version 2.0
Pull Request -
State: closed - Opened by sdrobert 10 months ago
#80 - Refactor docs and fix style
Pull Request -
State: closed - Opened by sdrobert about 1 year ago
#79 - LookupLanguageModel refactoring + reimplementing
Pull Request -
State: closed - Opened by sdrobert about 1 year ago
#78 - Standardize argument checking for modules
Pull Request -
State: closed - Opened by sdrobert over 1 year ago
#77 - Allow scientific notation log-probs in ARPA
Pull Request -
State: closed - Opened by sdrobert over 1 year ago
#76 - Getting docs working again + lightning submodule init docs
Pull Request -
State: closed - Opened by sdrobert over 1 year ago
#75 - TrainingStateController automatically syncs with stored history less often
Pull Request -
State: closed - Opened by sdrobert over 1 year ago
#74 - SCPC updates n' fixes
Pull Request -
State: closed - Opened by sdrobert over 1 year ago
#73 - SCPC-related updates
Pull Request -
State: closed - Opened by sdrobert over 1 year ago
#72 - get-torch-spect-data-dir-info should be faster
Issue -
State: closed - Opened by sdrobert over 1 year ago
- 1 comment
#71 - Connection reset bug in distributed TrainingStateController
Issue -
State: open - Opened by sdrobert almost 2 years ago
- 2 comments
#70 - TextGrid handling and chunking of SpectDataSet
Pull Request -
State: closed - Opened by sdrobert almost 2 years ago
#69 - A metric ton of work from a project
Pull Request -
State: closed - Opened by sdrobert almost 2 years ago
#68 - Changed call + semantics of BeamSearch, RandomWalk; added FillAfterEndOfSequence
Pull Request -
State: closed - Opened by sdrobert over 2 years ago
#67 - Added binomial_coefficient, enumerate_binary_sequences_with_cardinality, and SRSWOR enumerate_support
Pull Request -
State: closed - Opened by sdrobert over 2 years ago
#66 - Updated documentation + docstrings
Pull Request -
State: closed - Opened by sdrobert over 2 years ago
#65 - Clean up and a bit of renaming/moving
Pull Request -
State: closed - Opened by sdrobert over 2 years ago
#64 - Implement RandomWalk for SequentialLanguageModel
Pull Request -
State: closed - Opened by sdrobert over 2 years ago
#63 - Redid pydrobert.torch.estimators, added .distributions
Pull Request -
State: closed - Opened by sdrobert over 2 years ago
#62 - Moved functionality to functional + modules submodules; refactoring
Pull Request -
State: closed - Opened by sdrobert over 2 years ago
#61 - JIT trace doesn't work for language models
Issue -
State: open - Opened by sdrobert almost 3 years ago
#60 - Isolate JIT scripting issues pre-1.8.0
Issue -
State: open - Opened by sdrobert almost 3 years ago
#59 - Function wrappers and JIT compatibility for most of the package
Pull Request -
State: closed - Opened by sdrobert almost 3 years ago
#58 - Compatibility fixes + setup changes
Pull Request -
State: closed - Opened by sdrobert almost 3 years ago
#57 - CTC and Beam Search interfaces, decoding modules, tutorials
Pull Request -
State: closed - Opened by sdrobert almost 3 years ago
#56 - Fix threshold issues for TrainingStateController (#55) and small CI speed-up
Pull Request -
State: closed - Opened by sdrobert over 3 years ago
#55 - TrainingStateController threshold implementations are incorrect
Issue -
State: closed - Opened by sdrobert over 3 years ago
- 1 comment
#54 - pad_variable and RandomShift for padding variable-length batches
Pull Request -
State: closed - Opened by sdrobert over 3 years ago
#53 - Fixes involving error rates
Pull Request -
State: closed - Opened by sdrobert over 3 years ago
#52 - error_rates should be based on the number of mistakes, not the total cost
Issue -
State: closed - Opened by sdrobert over 3 years ago
- 1 comment
#51 - Fix logic error in compute-torch-token-data-dir-error-rates and tests
Pull Request -
State: closed - Opened by sdrobert over 3 years ago
- 1 comment
#50 - compute-torch-data-dir-error-rates computes wrong score
Issue -
State: closed - Opened by sdrobert over 3 years ago
#49 - CTM-related fixes/features
Pull Request -
State: closed - Opened by sdrobert over 3 years ago
#48 - Refactor for Python 3
Pull Request -
State: closed - Opened by sdrobert over 3 years ago
#47 - Quick bug fix and updated build scripts
Pull Request -
State: closed - Opened by sdrobert over 3 years ago
#46 - SpecAugment and related utilities
Pull Request -
State: closed - Opened by sdrobert almost 4 years ago
#45 - Handle raw features in ctm<->ref with floating point frame shift
Pull Request -
State: closed - Opened by sdrobert about 4 years ago
#44 - Fixes for pytorch 1.6 API
Pull Request -
State: closed - Opened by sdrobert about 4 years ago
#43 - Bug fixes related to recent text data functionality for pytorch 1.0.1
Pull Request -
State: closed - Opened by sdrobert over 4 years ago
#42 - Better handling of text data
Pull Request -
State: closed - Opened by sdrobert over 4 years ago
#41 - Handling unk/oov in transcripts
Pull Request -
State: closed - Opened by sdrobert over 4 years ago
#40 - Added more sos pre-padding to LookupLanguageModel
Pull Request -
State: closed - Opened by sdrobert almost 5 years ago
#39 - Changed nomenclature in LookupLanguageModel, parse_arpa
Pull Request -
State: closed - Opened by sdrobert almost 5 years ago
#38 - Allow for empty backoffs for non-final n-grams, workaround for torch bug, and flake8
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#37 - Handle empty backoffs for non-final n-grams in parse_arpa_lm
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#36 - Language Modeling functionality
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#35 - Moved loss functions from training to layers
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#34 - Added the --replace option to compute-torch-token-data-dir-error-rates
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#33 - remaining_epoch -> remaining_epochs
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#32 - Gradient estimators updates
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#31 - Fixes to training scripts
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#30 - Changed reduce_lr_factor default
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#29 - Fixed max_ref_class bug in get-torch-spect-data-dir-info
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#28 - Updates to TrainingStateController, including semantics bug fix
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#27 - Added TrainingStateController.continue_training, tests
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#26 - Proper handling of optuna sampling
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#25 - More docstring updates
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#24 - Optuna hook redesign for pydrobert.param.optuna
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#23 - Retailored some aspects of data loaders + data params
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#22 - error_rate bug fix + faster implementation
Pull Request -
State: closed - Opened by sdrobert about 5 years ago
#21 - Removed optimizer_to and better handling of TrainingStateController persistence
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#20 - Fix time_distributed_return when gamma == 0.0
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#19 - MinimumErrorRateLoss takes 2d reference seqs, test
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#18 - Added time_distributed_return, tests
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#17 - A number of changes that improve consistency and clarity
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#16 - Additional param sampling for optuna in pydrobert.torch.data
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#15 - Updates to TrainingStateController
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#14 - Added batch_first option to data loaders
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#13 - Simplify MinimumErrorRate + add utilities
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#12 - No dropout prob in TrainingStateParams
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#11 - Torch 1.2.0 compatibility
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#10 - Docs updates
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#9 - Hotfix for gradient estimators
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#8 - Sampling
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#7 - Updated docs
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#6 - Various attention layers
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#5 - Objectives, auxiliary functions, and metrics related to decoding
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#4 - Added ReadTheDocs
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#3 - Drastic revamp of data submodule
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#2 - Support for gradient estimators
Pull Request -
State: closed - Opened by sdrobert over 5 years ago
#1 - Travis CI build fixes and updates
Pull Request -
State: closed - Opened by sdrobert over 5 years ago