facebookresearch/TimeSformer
The official pytorch implementation of our paper "Is Space-Time Attention All You Need for Video Understanding?"
PythonNOASSERTION
Issues
- 1
training from scratch, loss decreased very slow
#137 opened by chrisx599 - 1
Training issue
#110 opened by RebelliousTT - 0
- 2
Training at epoch 1 end up with CUDA error and Assertion `t >= 0 && t < n_classes` failed
#115 opened by kct22aws - 0
Prediction per frame
#135 opened by DanaOsama - 2
What exactly does cfg.TASK mean？
#108 opened by mudexuyan - 2
Finetuning code can be improved
#133 opened by feemthan - 3
Accuracy of each category
#129 opened by subin2202 - 1
How to reason about the video with the demo
#131 opened by xqc-qc - 0
Ok
#132 opened by Pshk1310136 - 1
ImageNet-1K Models
#96 opened by owaisCS - 0
What can I do if I can't find the train.csv file?
#130 opened by xqc-qc - 1
Video preprocessing to apply for trained model
#122 opened by sakethattaluri - 3
Kinetics dataset preparation
#113 opened by kct22aws - 2
Modified AVA dataset problem
#99 opened by MitchellWei - 0
Train on AVA dataset
#128 opened by chenqqqqqqq - 1
In "divided_space_time", Why is the class token not used during temporal attention?
#125 opened by Backdrop9019 - 0
- 0
- 1
How to apply this model for regression
#121 opened by Taimoor-R - 1
Question about train time
#117 opened by shiyi-z - 1
How to train driving dataset on timesformer?
#119 opened by pgupta119 - 2
input clips
#86 opened by W4ngH4o - 2
Has anyone managed to reproduce the TimeSformer (the basic model) results on K400?
#94 opened by kami93 - 2
Kinetic dataset labels
#114 opened by kct22aws - 0
During training, why is optimizer step called after 32 iterations?
#120 opened by karthikpullalarevu - 0
Why is the output 400 instead of 768？
#118 opened by lamsongianm - 2
Inference code reg
#102 opened by Jeba-create - 0
How to load pretrained SlowFast Model?
#116 opened by HashmatShadab - 0
How to train only the head？
#112 opened by lws6885 - 0
RuntimeError: Sizes of tensors must match except in dimension 0. Got 192 and 12 (The offending index is 0)
#111 opened by 906070493 - 0
problem about default_cfgs
#109 opened by cser-liu - 1
Attention Rollout for TimeSformer
#106 opened by yiyixuxu - 1
How to set "find_unused_parameters=True"
#107 opened by ppriyank - 0
what is the video_id id the train val csv file
#105 opened by yihui8776 - 0
- 0
Training on custom dataset
#104 opened by Jeba-create - 0
- 1
RotaryEmbedding
#100 opened by xieanheng - 0
Pre-trained models on ImageNet1k and 21k
#101 opened by wp2333 - 1
batchnorm and visualization
#98 opened by HC-2016 - 1
CUDA error: CUBLAS_STATUS_EXECUTION_FAILED
#95 opened by ardofski - 2
K600 labels?
#87 opened by cnut1648 - 2
Difference between Space and Joint Space-Time
#93 opened by himidev - 3
Fine-tuning TimeSformer on a smaller dataset
#88 opened by mtv21 - 1
influence of depth?
#89 opened by fanbooo - 1
That is a suggestion.
#90 opened by happyBroJk - 0
How should i test this model
#91 opened by happyBroJk - 1
- 1
Pre-trained model output prediction different from testing output prediction
#84 opened by chrisdt1998