Move the descriptions out of TabbedInferface.

Fix the errors in the examples of url tab.

Make the option of model and resolution available to both image and url inference tab.

Make the option of model and resolution available to both image and url inference tab.

Add an additional tab for inference with url.

Fix a bug.

Add the option of the legacy general use weights without training on portrait seg data.

Update descriptions in the app head.

simplify-inference-logic (#4)

Add default value of radio.

Add General-lite as another option in the model selection.

Upgrade the weights loading method to avoid duplicated loading.

Upgrade the weights loading method to avoid duplicated loading.

Upgrade the weights loading method to avoid duplicated loading.

Upgrade the weights loading method to avoid duplicated loading.

Upgrade the weights loading method to avoid duplicated loading.

Upgrade the weights loading method to avoid duplicated loading.

Upgrade the weights loading method to avoid duplicated loading.

Upgrade the weights loading method to avoid duplicated loading.

Upgrade the weights loading method to avoid duplicated loading.

Add weights option to BiRefNet trained in all different settings.

Add weights option to BiRefNet trained in all different settings.

Add weights option to BiRefNet trained in all different settings.

Add weights option to BiRefNet trained in all different settings.

Update the model switch part.

Update the model switch part.

Add the weights option.

Add the weights option.

Update the description in the head of the main page, since this project can access the zeroGPU now.

Update app.py

Update app.py

Upgrade the space codes with deployed HF model.

Remove unnecessary weights.

Update weights.

Update weights.

Update app.py

Update app.py

Use new BiRefNet with massive training and IoU fine-tuning.

Set presion of torch to high for consistency with training and faster inference.

Update the model codes, including the previous inconsistencies.

Update the weights to the one for general use with massive training.

Fix app.py (#2)

Make the comparison effect of the results more obvious (#3)

ZeroGPU and ImageSlider (#1)

Update the approximated time cost from 500s to 40s. The long time cost may be caused by the unnecessary part of our_ref in inference.

Remove redundant part of our_ref in inference.

Adapt the resize to cv2 version to keep consistent with that used in training.

Update the well-trained weights.

Add the link to my github repo.

Add the resolution value for examples.