Open yaomilwaukee opened 1 year ago
Same issue here. Hopefully this is fixed soon :(
Same issue
is that same with ? https://github.com/danielgatis/rembg/releases/download/v0.0.0/u2net_cloth_seg.onnx
Same issue.
I forked and retrained the model for 4 epochs (~90k iterations) with the given options. Had to make some adjustments due to newer package versions.
I don't know if my resulting model is as good. But if you need some quick predictions with an available pretrained model, you can refer to my fork: https://github.com/enrilohm/cloth-segmentation
Same issue here, could not find any properly cloth segment pre-trained model. Is this link https://huggingface.co/maiti/cloth-segmentation/tree/main the pre-trained model for this.
Same issue
Hi @levindabhi , the link for the pretrained weights is broken, could you please re-upload it ? Thanks a lot !
Hi @belgacemi @chiragksharma @hastaluegoph @enrilohm @pakresi @dalphajw @yaomilwaukee
For pretrained model download with inference script for Alpha Image generation and Cloth segmentation, please check this https://github.com/wildoctopus/huggingface-cloth-segmentation
Have cleaned, modified the scripts to run on latest pytorch and other libraries and Its very simple to use as well. The script will automatically download the pretrained model. So no worry at all.
You guys can also check the live demo at huggingface - https://huggingface.co/spaces/wildoctopus/cloth-segmentation
If you guys like this, please drop a star to my repo.
Thanks :)
Thanks @wildoctopus for this I am currently using segformer_b2_clothes for segmenting the clothes https://huggingface.co/mattmdjaga/segformer_b2_clothes I made a space for it as well :https://huggingface.co/spaces/mrrandom123/mattmdjaga-segformer_b2_clothes let me know which is better U2net or segformers
@wildoctopus and @chiragksharma , you guys are the legends. To all other, I ran both and found that @chiragksharma 's can be called as semantic seg and @wildoctopus 's as instance seg (just a rough example)
Hi @wildoctopus The output gives the correct image but the image details are completely destroyed. for example, passing an image of a person wearing a white shirt gives the image of the shirt as desired but the color is red. Is there any way to retain those image details ?
Hi @wildoctopus The output gives the correct image but the image details are completely destroyed. for example, passing an image of a person wearing a white shirt gives the image of the shirt as desired but the color is red. Is there any way to retain those image details ?
Hi @AkashKumar7902 , can you please share that input output pair, and what exactly you want in output? The model will output the segmented image, corresponds to upper body, lower body and full body cloth. You can see some differences in segmentted color because of color variations from background. But these things can be corrected by finetuning the model on such images.
And by "retaining", i belive you want to extarct the outfits based on the segmentation map, and because of color differences in soem outputs, not able to retain the details. So for that will suggest you to extract Details corresponds to each segmented color, and see if it works.
if the above reply isnot what you are expecting than add some input output pairs you are obtaining. Then will be able to clear what you need.
Hi @belgacemi @chiragksharma @hastaluegoph @enrilohm @pakresi @dalphajw @yaomilwaukee
For pretrained model download with inference script for Alpha Image generation and Cloth segmentation, please check this https://github.com/wildoctopus/huggingface-cloth-segmentation
Have cleaned, modified the scripts to run on latest pytorch and other libraries and Its very simple to use as well. The script will automatically download the pretrained model. So no worry at all.
You guys can also check the live demo at huggingface - https://huggingface.co/spaces/wildoctopus/cloth-segmentation
If you guys like this, please drop a star to my repo.
Thanks :)
excellent!
I found it, https://drive.usercontent.google.com/download?id=11xTBALOeUkyuaK3l60CpkYHLTmv7k3dY&authuser=0 I hope this one is the official one. I found it from here
https://github.com/ozzy404/cloth-segmentation fixed links and colab
same as title, google drive link of pre-trained model doesn't work. Could you upload a new one pre-trained model file?