Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / sillsdev/machine.py issues and pull requests

#120 - small fixes

Pull Request - State: closed - Opened by johnml1135 29 days ago - 1 comment

#119 - small fix for NMT build job

Pull Request - State: closed - Opened by johnml1135 about 1 month ago - 5 comments

#118 - Port USFM code from Machine up to commit c93f8dc

Pull Request - State: closed - Opened by mshannon-sil about 1 month ago - 4 comments
Labels: development

#117 - Update learning rate for #116 and other small error

Pull Request - State: open - Opened by johnml1135 about 1 month ago - 3 comments

#116 - Use improved learning rate schedule when training NMT jobs

Issue - State: open - Opened by ddaspit about 1 month ago - 1 comment
Labels: enhancement

#115 - Port USFM code from Machine up to commit bf2b46d

Pull Request - State: closed - Opened by mshannon-sil about 2 months ago - 2 comments
Labels: enhancement

#114 - Alignment Job

Pull Request - State: closed - Opened by johnml1135 about 2 months ago - 26 comments

#113 - update shared_file_uri

Pull Request - State: closed - Opened by mshannon-sil 2 months ago - 2 comments

#112 - Change default shared_file_uri

Issue - State: closed - Opened by mshannon-sil 2 months ago

#111 - Port USFM code from Machine up to commit a9058ce

Pull Request - State: closed - Opened by mshannon-sil 3 months ago - 2 comments
Labels: enhancement

#110 - Scripture range parser no error on empty string

Pull Request - State: closed - Opened by Enkidu93 3 months ago - 1 comment

#109 - Use macos-12 GHA runner

Pull Request - State: closed - Opened by ddaspit 5 months ago - 1 comment

#108 - Build crashes with thot

Issue - State: closed - Opened by johnml1135 5 months ago - 1 comment
Labels: bug

#107 - Smt build job

Pull Request - State: closed - Opened by johnml1135 5 months ago - 44 comments

#106 - Parallel to https://github.com/sillsdev/machine/pull/187

Pull Request - State: closed - Opened by Enkidu93 6 months ago - 1 comment

#105 - fix normalization order in find_missing_characters()

Pull Request - State: closed - Opened by mshannon-sil 6 months ago - 1 comment
Labels: bug

#104 - normalize lines before getting charset

Issue - State: closed - Opened by mshannon-sil 6 months ago
Labels: bug

#103 - Update machine.py to reflect usfm changes in machine

Pull Request - State: closed - Opened by mshannon-sil 6 months ago - 2 comments
Labels: enhancement

#102 - Port Machine's USFM generation changes to machine.py

Issue - State: open - Opened by mshannon-sil 6 months ago - 1 comment
Labels: enhancement

#100 - And upgrade torch, numpy, pandas and accelerate as well as black, pyr…

Pull Request - State: closed - Opened by johnml1135 8 months ago - 2 comments

#99 - Add option to save the model during build job

Pull Request - State: closed - Opened by ddaspit 8 months ago - 13 comments

#98 - Refactor get_chapters to not reuse code and add additional tests

Pull Request - State: closed - Opened by isaac091 9 months ago - 1 comment

#97 - Clean up get_chapters code and expand tests

Pull Request - State: closed - Opened by isaac091 9 months ago

#96 - use sacremoses normalizer, ensure pretranslate.src.json and pretranslate.trg.json use same directory

Pull Request - State: closed - Opened by mshannon-sil 9 months ago - 2 comments
Labels: bug

#94 - Quotation marks are stripped out when using NLLB

Issue - State: closed - Opened by ddaspit 9 months ago
Labels: bug

#93 - Add test for default attributes in USFM

Pull Request - State: closed - Opened by ddaspit 9 months ago

#92 - swap out 600M param model for tiny for faster startup

Pull Request - State: closed - Opened by johnml1135 9 months ago - 1 comment

#91 - When an SFM file fails to parse report the full file path and if possible the line number.

Issue - State: closed - Opened by davidbaines over 1 year ago - 2 comments
Labels: enhancement, good first issue

#90 - Update error message to include filename

Pull Request - State: closed - Opened by isaac091 10 months ago - 2 comments

#89 - unsupported language code

Issue - State: open - Opened by johnml1135 10 months ago - 1 comment
Labels: documentation, deployment

#88 - Restrict build options to only update model hyperparameters

Pull Request - State: closed - Opened by ddaspit 10 months ago - 11 comments

#87 - Create more robust check for old vs new book selection syntax

Pull Request - State: closed - Opened by isaac091 10 months ago - 1 comment

#86 - Restrict build options to only update model hyperparameters

Issue - State: closed - Opened by ddaspit 10 months ago - 6 comments
Labels: bug

#85 - Add a "analyze data" job

Issue - State: closed - Opened by johnml1135 10 months ago - 1 comment
Labels: enhancement

#84 - Automatically run wildebeest

Issue - State: closed - Opened by johnml1135 10 months ago - 1 comment
Labels: enhancement

#83 - Support NMT training on USFM tokens for poetry etc.

Issue - State: open - Opened by johnml1135 10 months ago
Labels: enhancement

#82 - Upgrade machine.py production docker container to python 3.11

Pull Request - State: closed - Opened by mshannon-sil 10 months ago - 1 comment
Labels: enhancement

#81 - distill or prune model to save training time

Issue - State: open - Opened by johnml1135 10 months ago - 1 comment
Labels: enhancement

#80 - Update get_chapters to more similarly to get_books

Pull Request - State: closed - Opened by isaac091 10 months ago - 1 comment

#79 - Add SentencePiece tokenizer (with options)

Issue - State: open - Opened by johnml1135 10 months ago
Labels: enhancement

#78 - Add 2 more A100's

Issue - State: closed - Opened by johnml1135 10 months ago
Labels: deployment

#77 - Fixed off-by-one errors in _create_word_graph

Pull Request - State: closed - Opened by isaac091 10 months ago - 1 comment

#76 - Update Thot to 3.4.3

Pull Request - State: closed - Opened by ddaspit 10 months ago

#75 - Crash in FuzzyEditDistanceWordAlignmentMethod

Issue - State: closed - Opened by ddaspit 10 months ago
Labels: bug

#74 - Fix crash in FuzzyEditDistanceWordAlignmentMethod

Pull Request - State: closed - Opened by ddaspit 10 months ago - 1 comment

#73 - Reduce memory use through better attention?

Issue - State: open - Opened by johnml1135 10 months ago - 3 comments

#72 - Torch compile for faster inferencing?

Issue - State: closed - Opened by johnml1135 10 months ago - 3 comments

#71 - Accelerate - run our models distributed very easily?

Issue - State: open - Opened by johnml1135 10 months ago
Labels: enhancement, requirement

#70 - TF32 - we should enable this - or is it already done?

Issue - State: closed - Opened by johnml1135 10 months ago - 1 comment

#69 - Remove obsolete code from NmtModelFactory

Pull Request - State: closed - Opened by ddaspit 11 months ago

#68 - Correctly handle corrupted SMT models

Pull Request - State: closed - Opened by ddaspit 11 months ago - 1 comment

#67 - Upgrade to torch 2.1.1 and fix OOM exception handling

Issue - State: closed - Opened by johnml1135 11 months ago

#66 - map windows codepage 936 to python encoding cp936

Pull Request - State: closed - Opened by mshannon-sil 11 months ago - 1 comment
Labels: bug

#65 - LookupError: unknown encoding: gb2313

Issue - State: closed - Opened by mshannon-sil 11 months ago
Labels: bug

#64 - From pretrained normally?

Issue - State: open - Opened by johnml1135 11 months ago - 3 comments
Labels: ci

#63 - Better error messages for #62

Pull Request - State: closed - Opened by johnml1135 11 months ago - 4 comments

#62 - ValueError : 'grc_Cprt' is not a valid language code.

Issue - State: closed - Opened by pmachapman 11 months ago - 4 comments

#61 - Move get_books to parse.py and add get_chapters

Pull Request - State: closed - Opened by isaac091 11 months ago - 12 comments

#60 - convert book abbreviation in \id marker to uppercase

Pull Request - State: closed - Opened by mshannon-sil 11 months ago - 4 comments
Labels: bug

#59 - 3 letter abbreviation for bible books should always be uppercase

Issue - State: closed - Opened by mshannon-sil 11 months ago - 1 comment
Labels: bug

#58 - OOM error fixing

Pull Request - State: closed - Opened by johnml1135 11 months ago - 16 comments

#57 - handle stopped tasks without throwing error

Pull Request - State: closed - Opened by mshannon-sil 11 months ago
Labels: bug

#56 - Updating ClearML Task after stopped faults build

Issue - State: closed - Opened by Enkidu93 11 months ago - 9 comments
Labels: bug

#55 - enforce upper bound on max steps

Pull Request - State: closed - Opened by mshannon-sil 11 months ago - 1 comment
Labels: enhancement

#54 - Validation split option

Issue - State: open - Opened by johnml1135 11 months ago - 1 comment
Labels: enhancement

#53 - Build summary - for debugging/diagnostics

Issue - State: open - Opened by johnml1135 11 months ago - 8 comments
Labels: enhancement

#52 - Out of memory

Issue - State: closed - Opened by johnml1135 11 months ago - 16 comments
Labels: bug

#51 - max (max) steps - 50,000

Issue - State: closed - Opened by johnml1135 11 months ago - 2 comments
Labels: enhancement

#50 - Raise error when passing an invalid lang code to HuggingFaceNmtEngine

Pull Request - State: closed - Opened by ddaspit 11 months ago

#49 - Add TranslationSuggester and friends

Pull Request - State: closed - Opened by ddaspit 12 months ago - 4 comments

#48 - Fix path issue for loading from custom_normalizer directory

Pull Request - State: closed - Opened by mshannon-sil 12 months ago
Labels: bug

#47 - Unable to find custom_normalizer directory

Issue - State: closed - Opened by mshannon-sil 12 months ago
Labels: bug

#46 - Fix corpus count methods

Pull Request - State: closed - Opened by ddaspit 12 months ago

#45 - Set HuggingFaceNmtEngine to not truncate by default

Pull Request - State: closed - Opened by ddaspit 12 months ago - 3 comments

#44 - Non-NllbTokenizers (non-fast) aren't able to save and load new language codes

Issue - State: open - Opened by mshannon-sil 12 months ago - 4 comments
Labels: bug

#43 - Tokenizer

Pull Request - State: closed - Opened by mshannon-sil 12 months ago - 9 comments
Labels: enhancement

#42 - environment types

Pull Request - State: closed - Opened by johnml1135 12 months ago - 2 comments

#41 - Percent complete try2

Pull Request - State: closed - Opened by johnml1135 12 months ago - 7 comments

#40 - Out Of Memory - very long input lengths?

Issue - State: closed - Opened by johnml1135 12 months ago - 3 comments
Labels: bug

#39 - Suppport training on intra-verse USFM

Issue - State: closed - Opened by johnml1135 12 months ago - 1 comment
Labels: enhancement

#38 - add codecov

Pull Request - State: closed - Opened by mshannon-sil about 1 year ago - 1 comment
Labels: ci

#37 - custom message for type errors when parsing build options

Pull Request - State: closed - Opened by mshannon-sil about 1 year ago
Labels: bug

#36 - Config passing serval #45

Pull Request - State: closed - Opened by mshannon-sil about 1 year ago - 1 comment

#35 - Update python to 3.11

Issue - State: closed - Opened by mshannon-sil about 1 year ago
Labels: enhancement

#33 - Add codecov

Issue - State: closed - Opened by johnml1135 about 1 year ago - 5 comments
Labels: ci

#32 - Config passing serval #45

Pull Request - State: closed - Opened by mshannon-sil about 1 year ago - 9 comments
Labels: enhancement

#31 - Update regex

Pull Request - State: closed - Opened by ddaspit about 1 year ago - 2 comments

#30 - Add support for Python 3.11

Pull Request - State: closed - Opened by ddaspit about 1 year ago - 1 comment

#29 - catch all errors and mark clearml task as failed with status_reason

Pull Request - State: closed - Opened by mshannon-sil about 1 year ago - 5 comments
Labels: enhancement

#28 - build job global exception handling - log and propagate

Issue - State: closed - Opened by johnml1135 about 1 year ago - 4 comments
Labels: bug

#27 - Attach to running container

Issue - State: closed - Opened by johnml1135 about 1 year ago - 2 comments
Labels: development

#26 - Nonexistent status "stopping" in clearml_check_canceled()

Issue - State: closed - Opened by mshannon-sil about 1 year ago - 2 comments
Labels: bug, invalid

#25 - cache pytorch_model.bin

Issue - State: closed - Opened by johnml1135 about 1 year ago - 9 comments
Labels: enhancement

#24 - Docker fixes

Pull Request - State: closed - Opened by johnml1135 about 1 year ago

#23 - IndexError: Invalid key: 0 is out of bounds for size 0

Issue - State: closed - Opened by johnml1135 about 1 year ago - 25 comments
Labels: bug

#22 - Out of Memory when translating

Issue - State: closed - Opened by johnml1135 about 1 year ago - 1 comment

#21 - Proper error handling for very long segments

Issue - State: closed - Opened by johnml1135 about 1 year ago - 4 comments
Labels: bug

#20 - Update tokenizer to accept new characters for Huggingface models

Issue - State: closed - Opened by johnml1135 about 1 year ago - 1 comment
Labels: enhancement