Not getting sensible result!

#1
by alir3za97 - opened

I tried multiple images with different prompts, but I can't seem to get any correct mask predictions.
For example this is one of facebooks examples in the meta demo.
I tried many prompts like: "bus", "a photo of a bus", " a bus in street" but didn't get the desired output.
image.png

This comment has been hidden

Sorry. It was my fault.
There was a bug in image preprocessing and I fixed it.

스크린샷 2023-04-09 오전 11.33.44.png

this seems a bit strange,a face > a body
image.png

@metrosir
I checked several results on the latest commit:

스크린샷 2023-04-27 오전 11.44.41.png
스크린샷 2023-04-27 오전 11.43.42.png
스크린샷 2023-04-27 오전 11.43.05.png
스크린샷 2023-04-27 오전 11.42.36.png
스크린샷 2023-04-27 오전 11.41.56.png

Hi, I'm just exploring https://huggingface.co/spaces and found so many amazing things. Is there a way to implement these examples locally or use them with an API?

@arjuna-d3v you can git clone this

grafik.png

run pip3 install -r requirements.txt and then start it with gradio app.py, check for the localhost url in the terminal and open it in your browser. you will have the same app running locally now.

Sign up or log in to comment