facebookresearch/fairseq-lua

Issues

• Torchscript EncoderDecoderTransformer

  #147 opened 3 years ago by msaroufim
  0

• Can I use WordPiece tokenizer when I pretrain RobertaModel following the guideline?

  #149 opened 3 years ago by qwer4107
  1

• Cannot build model

  #148 opened 3 years ago by at3e
  0

• Problems in comparing parameters

  #146 opened 4 years ago by leonleeldc
  0

• [question] how to know use which arch parameters?

  #145 opened 4 years ago by HyejinWon
  4

• Inference/finetuning Blender in fairseq

  #144 opened 4 years ago by hckh
  2

• The gradient (Tensor.grad) of decoder weights is None

  #143 opened 4 years ago by NonvolatileMemory
  1

• Data preparation

  #130 opened 6 years ago by zwx8981
  2

• Instructions for fine-tuning pre-trained models

  #139 opened 5 years ago by y3nk0
  1

• Pretrained word embeddings for machine translation?

  #141 opened 5 years ago by darsh10
  1

• What is positional score per token?

  #140 opened 5 years ago by aj7tesh
  0

• Commands in quick start fail due to non-existence of S3 bucket

  #137 opened 5 years ago by bradheintz
  3

• error when load libbleu

  #138 opened 5 years ago by amelieyu
  0

• RuntimeError during training

  #136 opened 6 years ago by zhao1iang
  2

• How does the 1D convolution work?

  #135 opened 6 years ago by robrechtme
  2

• Typo in reference README

  #134 opened 6 years ago by robrechtme
  1

• Unable to Download Pretrained models

  #133 opened 6 years ago by nithya4
  2

• Encount CUDA error: out of memory when training multilingual translation task

  #132 opened 6 years ago by moonscar
  0

• generate.py RuntimeError: cublas runtime error : library not initialized at torch/lib/THC/THCGeneral.c

  #131 opened 6 years ago by shirakad
  2

• Segmentation fault during trainning

  #129 opened 6 years ago by mattiadg
  0

• Is Writing Prompts Dataset available in this repo ?

  #127 opened 6 years ago by GBLin5566
  1

• Cannot reverse order of translation

  #128 opened 6 years ago by andrewb-ms
  3

• sh: 1: ~/mosesdecoder/symal: Permission denied

  #126 opened 7 years ago by travel-go
  0

• Training with fconv model converges but not with blstm

  #125 opened 7 years ago by patrik-lambert
  0

• Averaging model parameters

  #124 opened 7 years ago by patrik-lambert
  2

• Pointers to distributed training

  #117 opened 7 years ago by posenhuang
  2

• create new class

  #99 opened 7 years ago by linhanxiao
  11

• Training error after use BPE

  #102 opened 7 years ago by happygirl123456
  2

• Generate from raw text

  #105 opened 7 years ago by nguyenlab
  1

• is fairseq support run as a service?

  #123 opened 7 years ago by huqitu
  1

• Generation with alignment dictionary

  #121 opened 7 years ago by patrik-lambert
  6

• Any method to reduce GPU memory usage?

  #122 opened 7 years ago by frankang
  1

• The problem of Chinese translation into English

  #120 opened 7 years ago by travel-go
  1

• Characters as input instead of words

  #119 opened 7 years ago by riteshpanjwani
  2

• Reproduce the result on WMT14 en-de

  #112 opened 7 years ago by Zrachel
  1

• I want to skip some epochs

  #118 opened 7 years ago by travel-go
  7

• the result of WMT en-de

  #116 opened 7 years ago by travel-go
  2

• out of

  #115 opened 7 years ago by travel-go
  3

• attention visualization of different decoder layers

  #114 opened 7 years ago by SuperWu090
  3

• Assertion in topk() failed during generation

  #113 opened 7 years ago by travel-go
  10

• How convolution exactly works in this network?

  #111 opened 7 years ago by aburkov
  4

• cannot reproduce result on wmt en-de

  #110 opened 7 years ago by Zrachel
  0

• regularization

  #109 opened 7 years ago by linhanxiao
  0

• loss regularization

  #107 opened 7 years ago by linhanxiao
  9

• out of memory at the end of epoch

  #108 opened 7 years ago by posenhuang
  2

• FairSeq model as a Server

  #104 opened 7 years ago by mdasadul
  6

• pytorch version

  #106 opened 7 years ago by linhanxiao
  2

• Does word embedding method have big impact of the training speed?

  #100 opened 7 years ago by nikefd
  3

• train the example model error: Segmentation fault

  #101 opened 7 years ago by Lingogo
  2

• question about network training

  #103 opened 7 years ago by linhanxiao
  3