ICNet-tensorflow

ICNet-tensorflow copied to clipboard

As you said, after opening the mean-var parameter, you need a big batch, otherwise the result is very bad. This is how I am now. My results are very bad....

AishuaiYao

Training on Cityscapes and Predict result worse

6

This is what I got, really bad result, the car can not segment out. I have trained with pretuned model about 4750 epochs.  Here is my predict codes, what...

lucasjinreal

same classification result with every pixel

1

When inference, why is my classification result of each pixel the same, all is No.5 classes？

AishuaiYao

Training on cityscapes dataset

3

Hi, I tried to run `train.py` to train the model from stretch on cityscapes dataset, and only commented the line that restores from pre-trained model and modified the path of...

junweiy

Update dependencies for who start from scratch.

Kyungpyo-Kim

tensorflow's version

1

thank you for providing the useful code! but when i run the demo file there are some error about the tensorflow's version and about the cuda cudnn. could you please...

leedoge

I use `train.py` to train on my relabel ade20k dataset (150 to 5 classes) but couldn't get good result. mIoU is always at maximum 57%. My training parameters: - Batch...

songshan0321

Please specific the environment config

2

Hi, hellochick! I run your code with meeting the problem as follow: `AttributeError: module 'tensorflow' has no attribute 'data' ` After I inquire this problem in the Google, I find...

Aaron9477

ValueError when using own dataset

2

Hi, I am trying to train ICNet on my own dataset using the cityscapes model as a starting point, but when I run train.py I receive the error: `ValueError: Dimension...

SpencerTrihus

tensorflow.python.framework.errors_impl.NotFoundError: Tensor name "conv1_1_3x3_s2/biases" not found in checkpoint files ./model/ade20k/model.ckpt-27150.index

2

Is the trained model not matching the network? @hellochick

RichardHansir