facebookresearch/AVT

Issues

• 50Salads annotations

  #47 opened a year ago by CeeZh
  0

• Problem when reproducing with Breakfast/50Salad data

  #43 opened a year ago by wasabipretzel
  2

• How to get attention visualizations for AVT?

  #46 opened 2 years ago by vineetparikh
  0

• EOFError: Ran out of input

  #45 opened 2 years ago by htcboy
  0

• Is there any alternative implementation for 'transformers.GPT2Model'?

  #44 opened 2 years ago by coolbay
  0

• -l seems to go into deadlock

  #31 opened 3 years ago by haofengac
  7

• EK55 primary meric report

  #42 opened 2 years ago by ofir1080
  2

• The input/output feature dimensions of Transformer Encoder and Causal Transformer Decoder?

  #41 opened 3 years ago by yxgz
  1

• Generate submitted file

  #39 opened 3 years ago by interestingzhuo
  3

• Loading pretrained ViT base as a backbone

  #38 opened 3 years ago by ofir1080
  3

• Long-term anticipation

  #40 opened 3 years ago by RodinIvan
  1

• Can I train AVT on RGB frames?

  #37 opened 3 years ago by forrestsz
  2

• Data preparation (EK)

  #36 opened 3 years ago by ofir1080
  2

• Use RULSTM feature

  #34 opened 3 years ago by sanketsans
  2

• Cannot train/test with video data

  #35 opened 3 years ago by CodyQ3
  8

• [Question] Video frame prediction without actions

  #24 opened 3 years ago by austinmw
  1

• Top-1 Accuracy EK100 RGB

  #33 opened 3 years ago by zerodecoder1
  1

• cant create video dataset

  #30 opened 3 years ago by sanketsans
  5

• Couple questions about classification loss

  #27 opened 3 years ago by zerodecoder1
  9

• Can't produce result for pre-trained TSN. Takes long time

  #29 opened 3 years ago by ttran1904
  13

• Doubts regarding the Causality

  #28 opened 3 years ago by Anirudh257
  2

• Unable to reproduce '01_ek100_avt.txt' val result

  #23 opened 3 years ago by zhoumumu
  7

• Job exiting automatically after a while

  #26 opened 3 years ago by Anirudh257
  0

• How to print the noun/verb class accuracies for EGTEA and Epic-Kitchens?

  #25 opened 3 years ago by Anirudh257
  2

• Regarding sampling method and strategy

  #19 opened 3 years ago by ofir1080
  4

• Number of GPUs Used During End-to-end Training

  #22 opened 3 years ago by okay-okay
  2

• Some questions about future_prediction.py

  #14 opened 3 years ago by okay-okay
  3

• Data Augmentation and End-to-end Training

  #15 opened 3 years ago by okay-okay
  3

• Env Create Conflicts

  #16 opened 3 years ago by GOZGOXGOYGO
  5

• Unable to initialize from the pretrained model

  #17 opened 3 years ago by TanviAgg
  1

• Recall / Accuracy Questions on Paper

  #21 opened 3 years ago by okay-okay
  1

• Reproduce models using avt-b on EK55

  #18 opened 3 years ago by ofir1080
  2

• About EK55.

  #20 opened 3 years ago by AllenXuuu
  1

• Data Augmentations Used on Training Data

  #6 opened 3 years ago by okay-okay
  2

• Usage of avg_last_n in future_prediction.py

  #13 opened 3 years ago by okay-okay
  1

• Questions about Paper Results

  #12 opened 3 years ago by okay-okay
  3

• fps of extracted irCSN feature

  #11 opened 3 years ago by AllenXuuu
  1

• Question about Object/Image Features

  #9 opened 3 years ago by okay-okay
  3

• EK100 Validation Using Only RGB Modality

  #10 opened 3 years ago by okay-okay
  1

• Unable to reproduce val results

  #8 opened 3 years ago by haofengac
  7

• Calculation for Mean/STD Statistics

  #7 opened 3 years ago by okay-okay
  1

• Plans to release EK-55 / EGTEA Gaze+ models?

  #5 opened 3 years ago by haofengac
  1