Ecosyste.ms: Issues

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

GitHub / prajdabre/yanmtt issues and pull requests

#66 - To avoid StopIteration error. Change from while-loop to for-loop.

Pull Request - State: open - Opened by hour 4 months ago

#64 - Use non-root install for sentencepiece

Pull Request - State: open - Opened by LoicGrobol 12 months ago

#63 - Disable shared sentencepiece libraries in installation instructions

Issue - State: open - Opened by LoicGrobol 12 months ago - 1 comment

#62 - Log final loss reduced across all processes instead of rank 0

Pull Request - State: open - Opened by jaygala24 about 1 year ago

#61 - add activation functions, rms norm and wandb logging

Pull Request - State: closed - Opened by jaygala24 over 1 year ago

#59 - Problem with __future__ annotation

Issue - State: open - Opened by yash-srivastava19 over 1 year ago - 1 comment

#58 - Pretrain Donut model

Issue - State: open - Opened by Wyzix33 over 1 year ago - 1 comment

#57 - Extending IndicBART or IndicBERT

Issue - State: open - Opened by singhakr almost 2 years ago - 7 comments

#56 - Alternative to installing sentencpiece

Issue - State: open - Opened by Sreyan88 almost 2 years ago - 11 comments

#54 - Error in BART Monolingual Pre-training.

Issue - State: closed - Opened by Ab26 almost 2 years ago - 5 comments

#53 - Support all optimizers and schedulers

Issue - State: open - Opened by prajdabre about 2 years ago
Labels: good first issue

#52 - Mixtures of denoisers

Issue - State: open - Opened by prajdabre about 2 years ago

#51 - Add post-norm to the model

Issue - State: open - Opened by prajdabre about 2 years ago
Labels: good first issue

#50 - Add PEP8 style guide checker workflow

Issue - State: open - Opened by divyanshugit about 2 years ago
Labels: good first issue

#48 - Display more information during training

Issue - State: open - Opened by prajdabre about 2 years ago
Labels: good first issue

#47 - Add support for latest version of transformers repo

Issue - State: open - Opened by prajdabre about 2 years ago
Labels: help wanted

#46 - CPU support

Issue - State: open - Opened by prajdabre about 2 years ago
Labels: good first issue

#45 - Binary executables for all python scripts

Issue - State: open - Opened by prajdabre about 2 years ago
Labels: documentation, enhancement

#44 - Improve examples

Issue - State: open - Opened by prajdabre about 2 years ago
Labels: documentation, help wanted

#43 - Improve documentation

Issue - State: open - Opened by prajdabre about 2 years ago
Labels: documentation

#42 - Adding WandB support

Issue - State: open - Opened by prajdabre about 2 years ago
Labels: good first issue

#41 - Cleaning the code for calling the model and computing loss

Issue - State: open - Opened by prajdabre about 2 years ago
Labels: documentation, good first issue

#40 - Some problem with loss

Issue - State: open - Opened by raullese about 2 years ago - 7 comments

#39 - some confused of <CUDA out of memory>

Issue - State: open - Opened by raullese about 2 years ago - 13 comments

#38 - Error when when trying to pretrain with other language extensions apart from hi

Issue - State: closed - Opened by raypretam about 2 years ago - 6 comments

#37 - Getting error when try to pre-train for three languages

Issue - State: open - Opened by Aniruddha-JU about 2 years ago - 15 comments

#36 - which pre-train model should we use for fine-tuning

Issue - State: open - Opened by Aniruddha-JU about 2 years ago - 2 comments

#34 - Getting error when pretraining with new languages sanskrit

Issue - State: open - Opened by Aniruddha-JU about 2 years ago - 12 comments

#33 - Some question of pretrain_nmt.py

Issue - State: open - Opened by raullese about 2 years ago - 1 comment

#32 - Exception: process 0 terminated with signal SIGSEGV

Issue - State: open - Opened by 1029694141 over 2 years ago - 8 comments

#31 - Pretraining Bart on Single corpus

Issue - State: closed - Opened by 1029694141 over 2 years ago - 2 comments

#30 - How to perform Inference using the fine-tuned model ?

Issue - State: closed - Opened by nikhilbyte over 2 years ago - 1 comment

#26 - Pre-training hangs

Issue - State: closed - Opened by jaspock over 2 years ago - 3 comments

#25 - GPU Consumption keeps on increasing

Issue - State: closed - Opened by nikhilbyte over 2 years ago - 1 comment

#24 - Trying to pretrain mBART model.

Issue - State: closed - Opened by nikhilbyte over 2 years ago - 4 comments

#23 - Broken package in requirements.txt

Issue - State: closed - Opened by ShivprasadSagare over 2 years ago - 1 comment

#22 - Invalid package reference in requirements.txt

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

#21 - dependency conflicts in requirements.txt

Issue - State: closed - Opened by ShivprasadSagare over 2 years ago - 5 comments

#20 - Using masked inputs at inference time

Issue - State: closed - Opened by jaspock over 2 years ago - 3 comments

#19 - Add xProphetNet model into this toolkit

Issue - State: open - Opened by koukoulala over 2 years ago - 3 comments

#18 - Train MBART Model: forward() got an unexpected keyword argument 'label_mask'

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

#17 - finetuning BART for text classification tasks as seq2seq

Issue - State: closed - Opened by GorkaUrbizu over 2 years ago - 3 comments

#16 - Revert "Dev"

Pull Request - State: closed - Opened by dipteshkanojia over 2 years ago

#15 - Revert "Dev"

Pull Request - State: closed - Opened by dipteshkanojia over 2 years ago

#14 - Dev

Pull Request - State: closed - Opened by dipteshkanojia over 2 years ago

#13 - Dev

Pull Request - State: closed - Opened by dipteshkanojia over 2 years ago

#12 - mBART embedding matrix prunning while finetuning on a single language

Issue - State: closed - Opened by GorkaUrbizu over 2 years ago - 3 comments

#11 - Issues at decode after finetuning mBART on a monolingual seq2seq task

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

#10 - Three custom languages and two tasks — is this a good place to start?

Issue - State: closed - Opened by jbmaxwell over 2 years ago - 4 comments

#8 - CUDA error while pre-training BART & how to use --hard_truncate_length

Issue - State: closed - Opened by GorkaUrbizu almost 3 years ago - 2 comments

#7 - Nan loss after 70K steps on the pre-training.

Issue - State: closed - Opened by GorkaUrbizu almost 3 years ago - 2 comments

#6 - Questions about BART pretraining hyperparameters.

Issue - State: closed - Opened by GorkaUrbizu almost 3 years ago - 2 comments

#5 - Can it be used to pre-train a BART model?

Issue - State: closed - Opened by Abigail61 almost 3 years ago - 1 comment

#4 - Tokenization issue when trying pretraining a monolingual model from scratch

Issue - State: closed - Opened by GorkaUrbizu almost 3 years ago - 3 comments

#3 - Continue training on pre-trained BART model

Issue - State: closed - Opened by pltrdy about 3 years ago - 1 comment

#2 - Tokenization issue with pretrained model

Issue - State: closed - Opened by pruksmhc about 3 years ago - 4 comments

#1 - AttributeError: 'Seq2SeqLMOutput' object has no attribute 'additional_lm_logits'

Issue - State: closed - Opened by pruksmhc about 3 years ago - 6 comments