Ecosyste.ms: Issues

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

GitHub / microsoft/MASS issues and pull requests

#50 - How to decide how many GPU is good for fine-tune ?

Issue - State: closed - Opened by XiaoqingNLP over 5 years ago

#50 - How to decide how many GPU is good for fine-tune ?

Issue - State: closed - Opened by XiaoqingNLP over 5 years ago

#47 - Can I initialize encoder parameters using bert pretrained model parameters?

Issue - State: closed - Opened by Wanjun0511 over 5 years ago - 1 comment

#47 - Can I initialize encoder parameters using bert pretrained model parameters?

Issue - State: closed - Opened by Wanjun0511 over 5 years ago - 1 comment

#46 - Conversational Response Generation

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

#46 - Conversational Response Generation

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

#45 - Illegal division by zero at multi-bleu.perl

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

#45 - Illegal division by zero at multi-bleu.perl

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

#44 - How to use multi-gpu to pretrain mass-fairseq model?

Issue - State: closed - Opened by ssxy00 over 5 years ago - 4 comments

#44 - How to use multi-gpu to pretrain mass-fairseq model?

Issue - State: closed - Opened by ssxy00 over 5 years ago - 4 comments

#43 - StopIteration Error with fairseq-interactive

Issue - State: open - Opened by magician-david over 5 years ago - 5 comments

#43 - StopIteration Error with fairseq-interactive

Issue - State: open - Opened by magician-david over 5 years ago - 5 comments

#42 - Rouge score Text Summarization

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

#42 - Rouge score Text Summarization

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

#41 - ZhEn pretraining model

Issue - State: open - Opened by Bournet over 5 years ago - 2 comments

#41 - ZhEn pretraining model

Issue - State: open - Opened by Bournet over 5 years ago - 2 comments

#40 - how to prepare the data to feed them into the fine-tuned model (Text-Sum)

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

#40 - how to prepare the data to feed them into the fine-tuned model (Text-Sum)

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

#39 - EN-only pre-trained model

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

#39 - EN-only pre-trained model

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

#38 - fp16 training

Issue - State: open - Opened by tdomhan over 5 years ago - 6 comments

#38 - fp16 training

Issue - State: open - Opened by tdomhan over 5 years ago - 6 comments

#37 - guide for MASS-fairseq Text Summarization preprocess and train

Issue - State: closed - Opened by BinWone over 5 years ago

#37 - guide for MASS-fairseq Text Summarization preprocess and train

Issue - State: closed - Opened by BinWone over 5 years ago

#36 - Use text-summarization model w/o tuning

Issue - State: closed - Opened by Arvedek over 5 years ago - 4 comments

#36 - Use text-summarization model w/o tuning

Issue - State: closed - Opened by Arvedek over 5 years ago - 4 comments

#35 - Results on CNN/DM dataset ?

Issue - State: closed - Opened by astariul over 5 years ago - 3 comments

#35 - Results on CNN/DM dataset ?

Issue - State: closed - Opened by astariul over 5 years ago - 3 comments

#34 - evaluator

Issue - State: closed - Opened by wengrx over 5 years ago - 4 comments

#33 - [MEMORY ERROR] RuntimeError: CUDA out of memory.

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

#32 - hyperparamters to reproduce paper result

Issue - State: open - Opened by xianxl over 5 years ago - 11 comments

#31 - How much data is used to train mass-fairseq model ?

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

#30 - size mismatch when reloading pretrained model

Issue - State: open - Opened by MichaelZhouwang over 5 years ago - 7 comments

#29 - [Data Ready] How to get /data/processed/en-fr ?

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

#27 - About Fine-tuning for Text Summarization

Issue - State: closed - Opened by dcn2020 over 5 years ago - 9 comments

#26 - The task of text summary task seems to be missing data ready

Issue - State: closed - Opened by lcl6679292 over 5 years ago - 3 comments

#25 - how to decode with the mass_ft_enfr_1024.pth

Issue - State: open - Opened by Bournet over 5 years ago - 6 comments

#24 - run get-data-gigaword.sh failed

Issue - State: closed - Opened by IreneZihuiLi over 5 years ago - 4 comments

#23 - Typo in MASS-fairseq

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

#22 - File Error When Finetuning Gigawords

Issue - State: closed - Opened by magic282 over 5 years ago - 9 comments

#21 - MASS Fairseq - A request

Issue - State: closed - Opened by sai-prasanna over 5 years ago - 6 comments

#20 - Quick question about pre-training with one sentence

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

#19 - About mask_sent of mass_step.Code and paper description seem to be inconsistent.

Issue - State: closed - Opened by miaodl over 5 years ago - 1 comment

#18 - How is evaluation done?

Issue - State: closed - Opened by deepaknlp over 5 years ago - 11 comments

#17 - func self.pred_layer.get_scores params error

Issue - State: closed - Opened by BinWone over 5 years ago - 1 comment

#14 - chinese model

Issue - State: open - Opened by joytianya over 5 years ago - 9 comments

#13 - Typo in shell script

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

#12 - Typo

Issue - State: closed - Opened by ndalton12 over 5 years ago - 1 comment

#11 - Baseline Implementations

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

#10 - How to use multi-gpu to pre-train the model?

Issue - State: closed - Opened by BinWone over 5 years ago - 3 comments

#9 - Would the en-zh pre-trained MASS model be public？

Issue - State: closed - Opened by chenweihua91 over 5 years ago - 2 comments

#8 - why is the data in folder hypotheses abnormal ?

Issue - State: closed - Opened by KelleyYin over 5 years ago - 4 comments

#7 - How can I make the data for text summarization task?

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

#6 - fast.cc: No such file or directory

Issue - State: closed - Opened by miaodl over 5 years ago - 1 comment

#5 - How can I reproduce the en-fr results of unsupervised NMT?

Issue - State: closed - Opened by KelleyYin over 5 years ago - 15 comments

#4 - Replicating en-fr UNMT with a smaller emb_dim

Issue - State: closed - Opened by magician-david over 5 years ago - 10 comments

#3 - Would the pre-trained model for text summarization be public also?

Issue - State: closed - Opened by IreneZihuiLi over 5 years ago - 7 comments

#2 - Why don't save the best model ?

Issue - State: closed - Opened by KelleyYin over 5 years ago - 1 comment

#1 - Can the pre-trained MASS model be directly used in XLM program ?

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