Zeqiang Lai
Zeqiang Lai
Add integration of raycast and wox, so that we can use any screenshot app to capture latex image and pass it to latexOCR server via pix2tex.api. BTW: It seems that...
I just could not make correspondence between the math and implementation of PDHG. It seems that we do not need to compute gradient when update x  https://github.com/odlgroup/odl/blob/0b088df8dc4621c68b9414c3deff9127f4c4f11d/odl/solvers/nonsmooth/primal_dual_hybrid_gradient.py#L282
For example, training log on SIDD 1. NAFNet-SIDD-width32 2. NAFNet-SIDD-width64 It would be helpful to compare the performance in the middle without entire training.
testdata: x_dc.mat sampling ratio is 0.2 parameters are set by default: ``` Rk = 7; weight = [0.015,0.015,0.015]; ``` log of last two iteration ``` **********************iter: 96 *********************** stopFormulaVal: 1.1146e-03...
See https://github.com/Zeqiang-Lai/DragGAN Have fun with it. https://github.com/XingangPan/DragGAN/assets/26198430/67d9467a-a9b3-4ec5-8777-60fe0258f429
> **Q: Expecting value: line 1 column 1 (char 0)** A: This might be caused by network proxy. Please refer to #50 for solutions. - #50
Feel free to use our workable implementation of DragGAN. https://github.com/Zeqiang-Lai/DragGAN https://github.com/Zeqiang-Lai/DragGAN/assets/26198430/f1516101-5667-4f73-9330-57fc45754283
- stylegan checkpoint pkl to pt conversion on demand - https://github.com/Zeqiang-Lai/DragGAN/pull/78 : CPU mode for GAN inversion
It get stucked while loading upfirdn2d_plugin. No response after 5 minutes
stabilityai/stable-diffusion-2-base DDIM config is not compatible with current DDIM implementation
### Describe the bug Assume 30 inference steps. There are two 925, and the number of inference steps is 31. tensor([958, 925, 925, 892, 859, 826, 793, 760, 727, 694,...