-
Hi, I like your project. When I tried to load the PyTorch model it returned error:
## My code:
```
checkpoint = torch.load("skin_u2net.pth", map_location='cpu', weights_only=False)
```
##…
-
wen running u2netp GPU not used,CPU running full
import torch
from PIL import Image
from config.conf import PROJECT_ROOT
from server.rembg.rembg import new_session, remove
rembg…
-
I have two models: u2net and u2netp, but node in my workflow dont see this models. Why?
This is node mix lab :
…
-
Why I am getting these errors despite having all my images and their corresponding binary files in their respective folders?
Learning the dataset itself...
Training Images Found:
[]
Number of …
-
Hi friend,
In your Readme.md, the information to check
Download Trained Models
Models can be downloaded [here](https://huggingface.co/skytnt/anime-seg)
however, when go to https://huggingface…
-
你好，我看了下训练命令和源代码。
第一行是训练了一个2dunet模型，但是在训练第二个模型，gln的时候，并没有使用到2dunet的权重，在test以及过程中，有加载模型的权重，但是并没有说明是哪个权重。
在我看来，应该是训练了两个模型，一个2dunet，然后一个后处理模型gln，预测以及使用需要两个模型结合，请问是我哪里理解错误了吗？
-
我用了U2Net，按照README里面的命令
./app_u2net --model=../../data/u2net/u2net.trt --size=320 --batch_size=1 --img=../../data/sailboat3.jpg --show --savePath
但这样只能推理一张图，把batch_size=8之后，应该怎么输入图像路径呢？
-
GPU: Nvidia 3090
CUDA: 11.2
ffmpeg: 5.0.1
model: u2net
-
First of all,thank you for your well organized and memory friendly code.
Could you kindly tell me how to use your code for training u2net on the custom dataset?
What are the steps?
Thanks in advanc…
-
Hey! Loved the paper.
Have you tried doing a comparison to U2Net? As for now, it's considered the state-of-the-art in SOD. I was wondering if you tested against it.