j-min/vl-t5 issues and pull requests

#36 - VL-T5 decoder all layers output how to get them ? size(batch_size, n_layers, sequence_length, hidden_dim)?

Issue - State: open - Opened by alice-cool over 1 year ago

#36 - VL-T5 decoder all layers output how to get them ? size(batch_size, n_layers, sequence_length, hidden_dim)?

Issue - State: open - Opened by alice-cool over 1 year ago

#35 - transformers version-up

Issue - State: open - Opened by j-min over 1 year ago
Labels: enhancement

#35 - transformers version-up

Issue - State: open - Opened by j-min over 1 year ago
Labels: enhancement

#34 - how to solve this error

Issue - State: open - Opened by its-athira over 1 year ago - 6 comments

#34 - how to solve this error

Issue - State: open - Opened by its-athira over 1 year ago - 6 comments

#33 - what's the difference between the 'evaluate' method and 'evaluate_raw' method?

Issue - State: open - Opened by CuddleSabe almost 2 years ago

#33 - what's the difference between the 'evaluate' method and 'evaluate_raw' method?

Issue - State: open - Opened by CuddleSabe almost 2 years ago

#32 - Text Generation Maximum Length

Issue - State: open - Opened by aboggust about 2 years ago - 1 comment

#32 - Text Generation Maximum Length

Issue - State: open - Opened by aboggust about 2 years ago - 1 comment

#31 - Can this model support Multilingual?

Issue - State: open - Opened by lonestar234028 about 2 years ago - 1 comment

#31 - Can this model support Multilingual?

Issue - State: open - Opened by lonestar234028 about 2 years ago - 1 comment

#30 - Extremely low zero-shot performance (0% acc on both val and test) on RefCOCOg

Issue - State: open - Opened by yiranyyu over 2 years ago - 1 comment

#30 - Extremely low zero-shot performance (0% acc on both val and test) on RefCOCOg

Issue - State: open - Opened by yiranyyu over 2 years ago - 1 comment

#29 - code wrong

Issue - State: closed - Opened by Su1008 over 2 years ago - 4 comments

#29 - code wrong

Issue - State: closed - Opened by Su1008 over 2 years ago - 4 comments

#28 - The model only pre-trained on the VG

Issue - State: open - Opened by zhangxi1997 over 2 years ago

#28 - The model only pre-trained on the VG

Issue - State: open - Opened by zhangxi1997 over 2 years ago

#27 - [Errno 32] Broken pipe

Issue - State: open - Opened by victorup over 2 years ago - 1 comment

#27 - [Errno 32] Broken pipe

Issue - State: open - Opened by victorup over 2 years ago - 1 comment

#26 - About the CPU memory leak

Issue - State: open - Opened by superhero-7 almost 3 years ago - 7 comments

#26 - About the CPU memory leak

Issue - State: open - Opened by superhero-7 almost 3 years ago - 7 comments

#25 - Some weights of VLT5VQA were not initialized from the model checkpoint at t5-base and are newly initialized

Issue - State: closed - Opened by aleSuglia almost 3 years ago - 1 comment

#24 - Update requirements.txt

Pull Request - State: closed - Opened by jnhwkim almost 3 years ago - 1 comment

#23 - Add Web Demo & Docker environment

Pull Request - State: closed - Opened by chenxwh almost 3 years ago - 1 comment

#22 - Error when finetuning RefCOCO with BART

Issue - State: closed - Opened by zuujhyt almost 3 years ago - 1 comment

#21 - a bug of VLT5TokenizerFast

Issue - State: open - Opened by Neo-Zhangjiajie almost 3 years ago - 1 comment

#20 - Considering adding refcoco inference example.

Issue - State: open - Opened by TaoHeyi almost 3 years ago

#19 - Another download link? I can't download vcr task "train_boxes36.h5"(22.4G)

Issue - State: open - Opened by alice-cool almost 3 years ago - 1 comment

#18 - WHY scripts GQA task batchsize of VL-T5 and VL-BART is not consistent of the paper?

Issue - State: closed - Opened by alice-cool almost 3 years ago - 1 comment

#17 - Could you please consider adding a license?

Issue - State: closed - Opened by zuujhyt about 3 years ago - 1 comment

#16 - Two gpu run gpa.py , encouner "[E ProcessGroupNCCL. cpp:341] Some NCCL operations have failed or timed out. Due to the asynchronous nature of CUDA kernels, "

Issue - State: closed - Opened by alice-cool about 3 years ago

#15 - forward function has encounter unexpected "head_mask"

Issue - State: closed - Opened by alice-cool about 3 years ago - 4 comments

#14 - About data feature files in GQA downstream tasks from where?

Issue - State: closed - Opened by alice-cool about 3 years ago

#13 - Low test accuracy on Refcocog

Issue - State: open - Opened by shizhediao about 3 years ago - 2 comments

#12 - No model.zero_grad()

Issue - State: closed - Opened by Richar-Du about 3 years ago - 2 comments

#11 - Error in provided Colab file

Issue - State: closed - Opened by dgknrsln about 3 years ago

#10 - Where do you override the generate method while inference (BART)

Issue - State: closed - Opened by zuujhyt about 3 years ago - 2 comments

#9 - Checkpoint on COCO captioning pretraining

Issue - State: closed - Opened by czy-orange about 3 years ago - 6 comments

#8 - 有谁能提供一下karpathy 的in domain 和out of domain 的分割数据集？

Issue - State: closed - Opened by alice-cool about 3 years ago

#7 - In domain and out of domain performance results from which datasets?

Issue - State: closed - Opened by alice-cool about 3 years ago

#6 - In domain and out of domain performance results from which datasets?

Issue - State: closed - Opened by alice-cool about 3 years ago - 1 comment

#5 - Captioning off-the-shelf

Issue - State: open - Opened by shrutijpalaskar about 3 years ago - 1 comment

#4 - Generator-based model work better than the classifier-based model?

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

#3 - How to complete Multi-task Finetuning

Issue - State: closed - Opened by Jiro-zhang over 3 years ago

#2 - Hyperparameter Tuning Strategies

Issue - State: open - Opened by shrutijpalaskar over 3 years ago - 1 comment

#1 - Inference on my own data?

Issue - State: open - Opened by puzzlecollector over 3 years ago - 21 comments

GitHub / j-min/vl-t5 issues and pull requests