microsoft/UniVL

Issues

• Non-Configurable GPU Count via Arguments

  #47 opened a year ago by willyfh
  0

• TypeError: bad operand type for unary -: 'list'

  #27 opened 3 years ago by jxrloveyou
  6

• How to only input text feature or video feature

  #40 opened 2 years ago by tingchihc
  2

• where to get transcript to generate youcookii_data.pickle

  #35 opened 3 years ago by zhaoying9105
  2

• Zero score (every output is None) on evaluation captioning with pretrained model

  #46 opened 2 years ago by Borntowarn
  1

• Weights from pretrained model not used in UniVL in evaluation. In EVALUATION, there is lack of visual_pytorch_model.bin, cross_pytorch_model.bin, decoder_pytorch_model.bin in visual-base, cross-base , decoder-base

  #5 opened 4 years ago by lokeaichirou
  12

• end-to-end video file captioning process

  #36 opened 3 years ago by mhyeonsoo
  3

• This repo is missing important files

  #33 opened 2 years ago by microsoft-github-policy-service
  1

• Issues about Freezing some additional layers instead of meanP in CLIP4Clip

  #43 opened 2 years ago by celestialxevermore
  2

• Estimate of zero-shot performance

  #45 opened 2 years ago by bpiyush
  1

• Error message (torch.distributed.elastic.multiprocessing.errors.ChildFailedError:)

  #44 opened 2 years ago by tingchihc
  0

• Is there a code for Finetune on CMU-MOSI here?

  #41 opened 2 years ago by sen0902
  1

• video only test for youcook

  #39 opened 2 years ago by mhyeonsoo
  2

• How can I create my video feature pickle

  #38 opened 2 years ago by tingchihc
  4

• feature & data shape

  #37 opened 3 years ago by mhyeonsoo
  6

• Unable to run video captioning code

  #34 opened 3 years ago by Davidyao99
  3

• Can you share your HowTo100M.csv file?

  #30 opened 3 years ago by ShinJQ
  3

• caption using features extracted from my raw video

  #11 opened 4 years ago by dawnlh
  6

• Pre-training acceleration using multi-machine distributed training

  #29 opened 3 years ago by mingtan2
  1

• Run Without Distributed

  #26 opened 3 years ago by Maddy12
  3

• How to run captioning task on my own video datasets?

  #28 opened 3 years ago by Kevinkaiyan
  1

• How to fine-tune with additional layers before UniVL?

  #25 opened 3 years ago by CrystalSixone
  2

• Questions on retrieval result and "Info: Weight doesn't exsits"

  #24 opened 3 years ago by HenryHZY
  4

• What's mean of the 'step_size=5' in modeling.py

  #23 opened 3 years ago by saicoco
  2

• CrossTask and COIN dataset code

  #22 opened 3 years ago by TXH-mercury
  1

• Joint loss in pretraining

  #21 opened 3 years ago by zhangliang-04
  1

• About auto mixed precision training

  #19 opened 3 years ago by zhangliang-04
  1

• How does the visual token come from?

  #20 opened 3 years ago by renmada
  1

• Hyper-parameter in pretraining

  #18 opened 3 years ago by zhangliang-04
  2

• About msrvtt retrieval results

  #17 opened 3 years ago by zhangliang-04
  1

• Captioning task clarification: video vs. video+text for captioning task

  #16 opened 3 years ago by tchang1997
  2

• The bert univl using is very different from Huggingfaces' (or pytorch's) bert

  #15 opened 3 years ago by butterluo
  0

• The program hangs when runs into parallel_apply() function in util.py

  #12 opened 3 years ago by butterluo
  7

• What's the role of the parameter coef_lr?

  #14 opened 4 years ago by forence
  1

• About multi-gpu loss calculation

  #13 opened 4 years ago by forence
  10

• caption my own video with provided pretrained model

  #10 opened 4 years ago by dawnlh
  8

• How should I set the value in youcookii_videos_features.pickle when fine-tuning with single transcript as input?

  #9 opened 4 years ago by lokeaichirou
  2

• Why is the fine-tuning performance much lower than benchmark in paper?

  #8 opened 4 years ago by lokeaichirou
  8

• Is the provided weights based on the pre-trained work on Howto100M dataset?

  #7 opened 4 years ago by lokeaichirou
  2

• RuntimeError: Default process group has not been initialized, please make sure to call init_process_group.

  #6 opened 4 years ago by lokeaichirou
  3

• CLip

  #4 opened 4 years ago by johnbager
  1

• When will you release the code of 'Action Step Localization' and 'Action Segmentation' tasks?

  #3 opened 4 years ago by butterluo
  1

• When will you release the pre-trained model？

  #2 opened 4 years ago by menggehe
  1

• Expected data format?

  #1 opened 4 years ago by mckinziebrandon
  3