issues
search
agrimgupta92
/
sgan
Code for "Social GAN: Socially Acceptable Trajectories with Generative Adversarial Networks", Gupta et al, CVPR 2018
MIT License
810
stars
259
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
How to reproduce S-LSTM with your code?
#80
sanmin0312
opened
4 years ago
2
error when running evaluate_model.py
#79
golnazhabibi3
opened
4 years ago
4
The reason for the embedding net
#78
Mirorrn
opened
4 years ago
0
Issue about training with datasets without downloading pretrained models
#77
YihaoChan
closed
4 years ago
0
Version of cuda and cudnn
#76
YihaoChan
closed
4 years ago
1
how can i train the social gan on a dataset that has been created by me
#75
ritwik18010
opened
4 years ago
9
Couldn't show the trajectory correctly
#74
lovelyyoshino
closed
4 years ago
0
What is the meaning of obs_traj,pred_traj_gt,pred_traj_fake etc ?
#73
ritwik18010
opened
4 years ago
1
questions on experiment of '4.2.2 Pooling in Action'
#72
CodeToPoem
opened
4 years ago
0
Extra line of code, doesn't serve any purpose
#71
sujithvemi
opened
4 years ago
0
Bump pillow from 5.1.0 to 6.2.0
#70
dependabot[bot]
closed
4 years ago
0
Question: Dataset Preprocess
#69
CHIH-CHIA
opened
4 years ago
4
GAN Loss implementation
#68
zhangpur
opened
4 years ago
0
M1: working on car data
#67
Chandrayee
closed
4 years ago
0
What's the mean of loss_mask_list and non_linear_ped ? More explain about dataset
#66
LittleFlyFish
opened
4 years ago
1
Bad prediction with pretrained model
#65
skpie
closed
4 years ago
0
pool_every_timestep in original paper?
#64
bysowhat
closed
4 years ago
1
Training loss does not change, and validation FDE error is super high
#63
cuihenggang
opened
4 years ago
6
Are the reported results on the validation set or the training set?
#62
abduallahmohamed
closed
4 years ago
0
Question about the datasets used for training
#61
cuihenggang
opened
4 years ago
4
Collisions: Pooling module may be ineffective
#60
davidglavas
opened
4 years ago
2
Eval results not the same as in the paper?
#59
Mirorrn
closed
4 years ago
0
typo in documentation?
#58
karttikeya
opened
4 years ago
0
question for data download
#57
xychen1015
opened
4 years ago
2
D_data_loss and G_discriminator_loss don't change
#56
agoodge
opened
5 years ago
15
Evaluation error for model without pooling module (KeyError : 'g_state')
#55
iamrjkunal
opened
5 years ago
3
Question about generator step loss
#54
Shunichi09
closed
5 years ago
2
ImportError: No module named sgan.data.loader
#53
aalzooke
opened
5 years ago
2
During training: "ValueError: could not convert string to float" because of delimiters
#52
agoodge
opened
5 years ago
5
Unreocgnized arguments for evaluate_model
#51
agoodge
closed
5 years ago
0
Doubt whether the pooling of SGAN-P really works in collision avoidance
#50
hunterbobo
opened
5 years ago
2
Training error TypeError: expected Tensor as element 1 in argument 0, but got tuple
#49
davidglavas
closed
5 years ago
1
a problem about the data download
#48
gwiplot
closed
5 years ago
0
differences between num_iterations and num_epochs
#47
w00zie
opened
5 years ago
0
The hyperparameters of 1V-1
#46
mingbocui
closed
5 years ago
0
According to the implementation details of the paper, the displacement error of the trained model is very large.
#45
oyjGithub
closed
5 years ago
0
The constant velocity model outperforms sgan?
#44
flclain
opened
5 years ago
3
Why did I encounter such a mistake?
#43
DENGZAIXU
opened
5 years ago
5
Why not using checkpoint['g_best_state'] in evaluate_model.py?
#42
mingbocui
closed
4 years ago
0
Why I could not get the same result with the paper?
#41
mingbocui
closed
5 years ago
8
predict for single trajectory
#40
saruvora
opened
5 years ago
0
Access frame number and ped_id
#39
saruvora
closed
5 years ago
1
Usages of loss_mask, linear_ped, Non_linear_ped in the code
#38
ybpaopao
closed
5 years ago
1
Training error: the GPU program failed
#37
saruvora
closed
5 years ago
1
stanford dataset
#36
saruvora
opened
5 years ago
10
About training and evaluation: in case target objects do not exist in 20 frames.
#35
himidev
closed
5 years ago
2
y_fake = torch.zeros_like(scores_fake) * random.uniform(0, 0.3)
#34
mingbocui
closed
5 years ago
1
there is a problem with the data download
#33
dlmuth
opened
5 years ago
5
Question: Error with shape
#32
saruvora
closed
5 years ago
3
Implementation of S-LSTM
#31
trungmanhhuynh
closed
5 years ago
2
Previous
Next