nv-tlabs / steal Goto Github PK

ERROR: Could not find a version that satisfies the requirement opencv-python==3.4.0.12 (from -r requirements.txt (line 3)) (from versions: 3.4.2.16, 3.4.2.17, 3.4.3.18, 3.4.4.19, 3.4.5.20, 4.0.0.21, 4.0.1.23, 4.0.1.24, 4.1.0.25)
ERROR: No matching distribution found for opencv-python==3.4.0.12 (from -r requirements.txt (line 3))

Hi there,
Thanks for the excellent work. I’m new to CV and machine learning in general and was wondering if it is possible to run this code if CUDA is not available, I.e. using cpu only or AMD gpu. Thanks!

Hi, thanks for your great work. I was curious about how the ground truth normal direction on boundaries is generated. It will be appreciated if you could share the implementation. Thank you.

when I use inference_cityscapes.py , it seems it will output different class into different images separately. I want to know how to get a total result like this in the paper?

HI, I am using the "pip install -r requirements.txt" for dependencies installation, but got following error. I have using ubantu with python 3.8.

ERROR: Could not find a version that satisfies the requirement torch==0.4.0 (from versions: 1.4.0, 1.5.0, 1.5.1, 1.6.0, 1.7.0, 1.7.1, 1.8.0, 1.8.1, 1.9.0, 1.9.1, 1.10.0, 1.10.1, 1.10.2, 1.11.0, 1.12.0, 1.12.1)
ERROR: No matching distribution found for torch==0.4.0

Please suggest a solution.

When I run the inference.py using your provided model and sdb dataset. the output images are almost dark and a few images have clear and thinning edges. I do not know why. Thank you for your reply!

I am running inference models, and the outputs are blank images.

(venv) (base) suryadi@xtal:/media/suryadi/DATA/learn/STEAL$ python inference_sbd.py \
>     --root_dir_val= ./data/sbd/data_aug/\
>     --flist_val= ./data/sbd/data_aug/val_list.txt\
>     --output_folder=./output/sbd/ \
>     --ckpt=./checkpoints/sbd/model_checkpoint.pt\
> 
usage: inference_sbd.py [-h] [--root_dir_val ROOT_DIR_VAL]
                        [--flist_val FLIST_VAL] [--ckpt CKPT]
                        [--output_folder OUTPUT_FOLDER] [--dataset DATASET]
                        [--n_classes N_CLASSES]
inference_sbd.py: error: unrecognized arguments: ./data/sbd/data_aug/ ./data/sbd/data_aug/val_list.txt

Dear David,

Where to download data?
Is it correct from http://www.eecs.berkeley.edu/Research/Projects/CS/vision/grouping/semantic_contours/benchmark.tgz ?

Thank you very much in advance.

Warmest Regards,
Suryadi

Thx for your work and releasing the test code.
How can I rewrite the direction loss if I didn't install matlab on my computer? Are there any other ways to calculate normal directions of the edge pixels?
Looking forward for your reply.

Hello.

Can you provide the scripts for the benchmarks on SBD re-annotated an Cityscapes? In fact, I took your pre-trained models and do the benchmakrs using the scripts from: https://github.com/Chrisding/seal, but I cannot get the same results as in your paper.

Thank you very much.

Thanks for the excellent work, but when will you release the training code?

Hello, thank you for your awesome work.

In the README.md file, the instruction to install the requirements has a typo. It should be:

pip install -r requirements.txt

Hello, thank you for your excellent work. I've read your paper and i wonder whether you plan to release the code of training, especially the computation details of NMS loss and direction loss. Thank you.

Hi! I just noticed that there is a small typo in your readme:
pip install -r requirments.txt

I want to train it on another dataset, thank you very much!