airsplay/lxmert

Issues

• Docker extract features error

  #121 opened 3 months ago by cmcv2022
  0

• FileNotFoundError: [Errno 2] No such file or directory: '/hy-tmp/lxmert/data/vqa/valid.json'

  #120 opened 4 months ago by nameyun
  0

• Corrupt VQA dataset from the source provided

  #119 opened 6 months ago by omm-prakash
  0

• Error in fine-tuning GQA for large datasets

  #118 opened 8 months ago by zilxv
  0

• Labels

  #116 opened 2 years ago by 25icecreamflavors
  0

• VQA dataset - how to calculate the label probability

  #87 opened 4 years ago by zixuwang1996
  2

• how to export onnx or tensorrt model when I using lxrt.modeling.LXRTModel

  #115 opened 2 years ago by aeeeeeep
  0

• Use of --qaSets parameter

  #99 opened 3 years ago by aurooj
  1

• training killed while loading image features

  #98 opened 3 years ago by Youngju-Na
  4

• how to know the object name from obj id in mscoco_imgfeat？

  #114 opened 2 years ago by xixihawokao
  0

• filtering raw dataset in to get tiny dataset

  #113 opened 3 years ago by samarIbrahem
  0

• Pre-training mission error

  #112 opened 3 years ago by MaxwellZYD
  0

• Fine-tuning VQA on multiple gpus

  #111 opened 3 years ago by simplelifetime
  0

• Get predictions for batches

  #110 opened 3 years ago by 25icecreamflavors
  0

• The URLs are not accessable

  #108 opened 3 years ago by BierOne
  12

• binascii.Error: Incorrect padding

  #101 opened 3 years ago by alice-cool
  1

• No problem

  #109 opened 3 years ago by alice-cool
  0

• Split .tsv files

  #79 opened 4 years ago by yikuan8
  11

• the result of Lxmert on vqa2.0 does not match the paper?

  #107 opened 3 years ago by hackerchenzhuo
  0

• validation score of VQA

  #106 opened 3 years ago by Yucheng-Han
  2

• Relation between `lxrt.entry.LXRTEncoder` and `lxrt.modeling.LXRTEncoder`.

  #92 opened 4 years ago by yezhengli-Mr9
  2

• why use tsv?

  #105 opened 3 years ago by Yucheng-Han
  0

• Training is stuck

  #104 opened 3 years ago by lhlm1994
  0

• Why having a self-attn after cross_attn

  #102 opened 3 years ago by LeungWaiHo
  1

• np.all(feature == 0) is True, why

  #103 opened 3 years ago by Yinhance
  0

• finetun VQA

  #100 opened 3 years ago by 1144181135
  0

• question on faster r-cnn feature extraction

  #97 opened 3 years ago by theaitetus
  0

• Bad performance of NLVR2.

  #84 opened 4 years ago by yangxuntu
  18

• Some question about nlvr

  #96 opened 3 years ago by haoopan
  0

• How to extract gqa_testdev_obj36.tsv feature?

  #95 opened 4 years ago by yangxuntu
  0

• Get embeddings of language encoder and Object-Relationship Encoder?

  #94 opened 4 years ago by yezhengli-Mr9
  2

• Is self.mode recommended for further debugging?

  #93 opened 4 years ago by yezhengli-Mr9
  0

• visual_attention_mask is set to None during pre-training?

  #91 opened 4 years ago by zhmd
  0

• How to get the language-only result?

  #82 opened 4 years ago by weiwenlan
  1

• cpu-only caffe version of extract_nlvr2_image.py

  #89 opened 4 years ago by yezhengli-Mr9
  3

• object/attr prediction not masked

  #90 opened 4 years ago by zhmd
  1

• Using LXMERT for image captioning

  #86 opened 4 years ago by freeIsa
  3

• In eval mode, the values are equal.

  #88 opened 4 years ago by Tramac
  0

• Pretrained model that wasn't pretrained on GQA

  #85 opened 4 years ago by yonatanbitton
  1

• Best pretrained model

  #83 opened 4 years ago by violetteshev
  0

• Not able to reproduce GQA(Train + BERT) 56.2

  #80 opened 4 years ago by yix081
  6

• Unable to overfit on tiny dataset (pre-training)

  #81 opened 4 years ago by hila-chefer
  1

• relationship between BatchSize and LR?

  #77 opened 4 years ago by XChuanLee
  3

• Freeze the model

  #78 opened 4 years ago by titaiwangms
  1

• PreTrain + BERT

  #76 opened 4 years ago by Muennighoff
  0

• VQA dataset

  #75 opened 4 years ago by deepaknlp
  2

• usage of tsv file (image feature extracted)

  #74 opened 4 years ago by coldtomatojuice
  1

• Class attribute of image features

  #73 opened 4 years ago by Kerry-zzx
  0

• Can I use -load_lxmert instead of -load_lxmert_qa when fine-tuning the VQA model?

  #72 opened 4 years ago by yikuan8
  1

• VQA: Split-wise result during each epoch

  #71 opened 4 years ago by gkv91
  0