-
Hello @GeorgeSeif!
Thanks for your work. I'm have a question about BiSeNet. Tried to train with default settings next configuration:
Dataset: CamVid
Arch: resnet50 + BiSenet
Image size: 512x51…
-
```
python train.py --name clean_10000 --dataroot ./datasets/clean_10000/ --no_instance --batchSize 5 --label_nc 0 --resize_or_crop none
```
Error
```
File "train.py", line 87, in
errors = {k…
-
I checked Camvid image dataset. I am trying to create annotated image just like Camvid annotated image but failed. Please suggest any tool or algorithm for the same. Please suggest can I use binary im…
-
At
lr_find(learn)
cuda runtime error (59) : device-side assert triggered at /opt/conda/conda-bld/pytorch-nightly_1540500315858/work/aten/src/THC/generic/THCTensorCopy.cpp:20
Such an error do…
-
Hi,
This model is suitable for binary segmentation. What if I want to do multi-label segmentation?
Best,
ghost updated
5 years ago
-
Hi GeorgeSeif,
Thanks for your implementation of so many SOTA segmentation networks.
I am interested in the RefineNet which works quite well on many datasets.
(1)I notice that the original pa…
-
@meetshah1995 @josephreisinger
The first epoch is normal:
Epoch [1/100] Loss: 3.4047
Epoch [1/100] Loss: 2.2006
Epoch [1/100] Loss: 1.7521
Epoch [1/100] Loss: 1.9977
Epoch [1/100] Loss: 1.9035
…
-
**Describe the bug**
**To Reproduce**
fastai version: 1.0.36.dev0
git pull to latest fastai
run lesson3-camvid.ipynb
error just after creating unet_learner on
`lr_find(learn)`
```
/…
-
### Information
## Please specify the following information when submitting an issue:
- **What are your command line arguments?:**
--dataset
CamVid
--model
RefineNet
--frontend
ResNet50
- **H…
-
Hi Fisher
Thanks for that great repository! Could you still tell me, how to convert the color images to label images, where each pixel has an ID that represents the ground truth label? Is there alrea…