hellochick PSPNet-tensorflow issues

hellochick / PSPNet-tensorflow

TensorFlow-based implementation of "Pyramid Scene Parsing Network".

326 stars 123 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

The Google Drive link for the checkpoint is not available

#69 Xposeder closed 2 weeks ago
2
small size

#68 mxkai opened 5 years ago
1
Error loading custom trained pspnet model file

#67 lakshmankanakala opened 5 years ago
0
feed_dict img in order to sequential sess.run() with multiple images

#66 kiitosu closed 5 years ago
0
Having trouble training

#65 AmeetR opened 5 years ago
2
How to get inference mask image , in which RGB values are correlating to cityscapes original label ids?

#64 LbinB opened 5 years ago
0
Is there anyone retrained the model successfully and get good result as this repo described?

#63 zhulf0804 closed 5 years ago
1
inference result with model trained with train.py is not good and model size is almost twise bigger then pretrained model

#62 kiitosu opened 5 years ago
3
Whether Extended Coarse DATASET

#61 Nedcome opened 5 years ago
0
How to get (class) label IDs or train IDs for evaluation with cityscapes-scripts

#60 ga84 opened 5 years ago
1
How to predict all images in a directory without loading model every time?

#59 TongJiayan opened 5 years ago
1
In train.py how to do validation as well as training

#58 XuhongLuo opened 5 years ago
0
No checkpoint file found

#57 ga84 closed 5 years ago
10
Training image size

#56 ghost opened 5 years ago
0
training from scratch

#55 mishaelran opened 5 years ago
1
PACC instead of mIoU

#54 mishaelran opened 5 years ago
1
using data augmentation decreases the resolution

#53 AyaMohamedS opened 5 years ago
0
i trained the new model,and modified the Num_class to 3,original 19, however, i got the results img fullfilled with pink color after i ran the inference, i don't know why?

#52 sainttelant opened 5 years ago
10
I wanna test my own picture, however ,when I use python inference.py --img-path=./input/indoor -1.jpg --dataset ade20k, it have some wrong ,I really don't know why?

#51 Kedron007 opened 5 years ago
2
ignoring label 0 in ade20k doesn't give the same labeling as the check point results

#50 AyaMohamedS closed 5 years ago
2
unexpected segmentation after training

#49 AyaMohamedS closed 5 years ago
1
tuning with selected images from ade20k

#48 AyaMohamedS closed 5 years ago
3
When the network reaches convergence, The loss has been around 2.0

#47 Hukongtao opened 5 years ago
14
The learning rate for different part of parameters

#46 YangWangsky closed 5 years ago
4
Resolution in evaluation

#45 Leequens opened 5 years ago
0
InvalidArgumentError (see above for traceback): assertion failed:

#44 Leequens closed 5 years ago
2
training error

#43 strand2013 opened 5 years ago
6
Initial model

#42 FengLoveBella opened 6 years ago
2
The auxiliary loss

#41 Ariel-JUAN opened 6 years ago
1
Evaluation failed with cityscapes

#40 sunggukcha closed 6 years ago
2
Issues with how to train the model from scratch

#39 zhiwenfan opened 6 years ago
1
Performance on VOC

#38 LongLong-Jing opened 6 years ago
0
Issue with the tutorial (https://modeldepot.io/hellochick/pspnet) - shapeless image

#37 PascalWB opened 6 years ago
1
Steps to training in different classes

#36 shadydiaa opened 6 years ago
0
Issue freezing graph

#35 mdlockyer opened 6 years ago
1
Training crashes on Cityscapes

#34 kshitijagrwl opened 6 years ago
2
What all inputs needed to finetune on a new dataset?

#33 314rated opened 6 years ago
0
How to better do batch inference on multiple images

#32 jin-zhe opened 6 years ago
4
Getting segmentation mask and class name predictions for ADE20k inference

#31 jin-zhe closed 6 years ago
3
Difference in Color Mapping to Label Id

#30 MichaelX99 closed 6 years ago
2
Fine tuning on different dataset

#29 Soumyabrata opened 6 years ago
1
Gray prediction image before encode

#28 shadydiaa opened 6 years ago
1
Can you train on the PASCAL dataset?

#27 YuHuangjie opened 6 years ago
0
Ckpt file for Cityscapes doesn't include all params to train

#26 anilarmagan closed 6 years ago
1
evaluation mIoU: 0.00404 on cityscapes val dataset

#25 amiltonwong opened 6 years ago
7
Pls share the trained models

#24 amiltonwong closed 6 years ago
1
bad inference result

#23 amiltonwong closed 6 years ago
3
Can you share your training procedure

#22 shipengai opened 6 years ago
1
how to train from scratch on ADE20K dataset

#21 shipengai opened 6 years ago
0
eval_flip

#20 FengLoveBella opened 6 years ago
0