hackkhai

[General Issue][Matting] Performance of MODNet is poor when trained on 1024x1024

8

Thanks for your issue. To help us solve the issue better, please provide following information： 1. PaddleSeg version: 2.6 2. PaddlePaddle version:2.3 3. Operation system: Linux 4. Python version:Python3.7 5....

Performance of MODNet is poor when trained on 1024x1024

8

Hey, I read your paper and its awesome!, I trained it on 512x512 resolution, it is really good, but when i trained the same data on higher resolution like 1024x1024...

So, I really love MGmatting's work on guidance based matting, But i found this weird matting issue in checkered pattern shirts though the trimap had only grey area around the...

Timeout for older cog docker execution locally using cog predict

1

Older dockers are stuck after pip installation and throws a timeout especially for the below one `cog predict r8.im/google-research/maxim@sha256:494ca4d578293b4b93945115601b6a38190519da18467556ca223d219c3af9f9 -i 'image="https://replicate.delivery/mgxm/6707a57f-4957-4047-b020-2160aed1d27a/1fromGOPR0950.png"' -i 'model="Image Deblurring (GoPro)"'` Error: `Building wheels for collected...

train_dreambooth_lora_flux_advanced ValueError: unexpected save model: <class 'transformers.models.t5.modeling_t5.T5EncoderModel'>

1

### Describe the bug Was exploring train_dreambooth_lora_flux_advanced.py from the advanced diffusion training examples. When I used --enable_t5_ti was getting the error ValueError: unexpected save model: ### Reproduction export MODEL_NAME="black-forest-labs/FLUX.1-dev" export...