SegFormer
SegFormer copied to clipboard
NVlabs
Official PyTorch implementation of SegFormer
error: OSError: EncoderDecoder: D:\SegFormer-master\pretrained\mit_b0.pth is not a checkpoint file how to solve it?
Hello dear author, in the paper to the SegFormer work it was referenced, that SegFormer starts with "first dividing input into patches of size 4 × 4, which is smaller...
Hi! Thanks for your excellent work, I want to train Segformer using imagenet, can you release the code for pre-training using imagenet ^ _ ^
As I know, `pytorch2onnx` from `mmseg` doesn't work correctly with `SegFormer` model. I have converted this, and accuracy is really bad (like random). So I have tried to export this...
I followed your prompts, but when I run the training code I get the following prompts, why is this, thanks! KeyError: "EncoderDecoder: 'mit_b1 is not in the models registry'"
For those having issues running the repo: Make sure that the versions of `pytorch` and `mmcv` exactly match. Furthermore, you also need to install `IPython` and `attrs` for the program...
Hi. Thanks for your Code and great community! I am training a custom dataset. The training goes fine, but when i do the inference, the result per class returns nan...
I used the following command for visualization, but got the following error. """ python demo/image_demo.py data/fp_grayscale/fp/images/validation/389808-2model.jpg local_configs/segformer/B4/segformer.b4.512x512.fp.40k_single-gpu.py work_dirs/segformer.b4.512x512.fp.40k_single-gpu/iter_26000.pth --device cuda:0 --palette fp Use load_from_local loader Traceback (most recent call last):...
[https://github.com/NVlabs/SegFormer/blob/master/docker/Dockerfile#:~:text=RUN%20git%20clone%20https%3A//github.com/open%2Dmmlab/mmsegmenation.git%20/mmsegmentation](url)
Why does the final output of the decoder scale down four times spatially, instead of being consistent with the input?
