https://github.com/kimrass/sdedit

PyTorch implementation of 'SDEdit' (Meng et al., 2021) from scratch and applying it to 'DDPM' on CelebA at 64 × 64
https://github.com/kimrass/sdedit

celeba ddpm image-quantization implementation median-filter pytorch sdedit

Last synced: 11 months ago
JSON representation

PyTorch implementation of 'SDEdit' (Meng et al., 2021) from scratch and applying it to 'DDPM' on CelebA at 64 × 64

• Host: GitHub
• URL: https://github.com/kimrass/sdedit
• Owner: KimRass
• Created: 2024-03-07T03:12:32.000Z (almost 2 years ago)
• Default Branch: main
• Last Pushed: 2024-05-30T08:41:52.000Z (over 1 year ago)
• Last Synced: 2024-05-30T10:18:07.162Z (over 1 year ago)
• Topics: celeba, ddpm, image-quantization, implementation, median-filter, pytorch, sdedit
• Language: Python
• Homepage:
• Size: 102 MB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# 1. Pretrained Models

- Download 'ddpm_celeba_64×64.pth' from [DDPM](https://github.com/KimRass/DDPM)

# 2. Samples

## 1) From Simulated Stroke

| `mode="from_sim_stroke"`,
`interm_time=0.30`, `ref_idx=132` | `mode="from_sim_stroke"`,
`interm_time=0.30`, `ref_idx=135` |

|:-:|:-:|

|  |  |

| `mode="from_sim_stroke"`,
`interm_time=0.45`, `ref_idx=132` | `mode="from_sim_stroke"`,
`interm_time=0.45`, `ref_idx=135` |

|:-:|:-:|

|  |  |

| `mode="from_sim_stroke"`,
`interm_time=0.60`, `ref_idx=132` | `mode="from_sim_stroke"`,
`interm_time=0.60`, `ref_idx=135` |

|:-:|:-:|

|  |  |

# 3. Theoretical Backgrounds

$$\mathbf{x}(t) = \alpha(t)\mathbf{x}(0) + \sigma(t)\mathbf{z}, \mathbf{z} \sim \mathcal{N}(\mathbf{0}, \mathbf{I})$$

# 4. To-Dos

- [x] Stroke input simulation.

- [x] Sampling from stroke.

- [ ] Total repeats.

- [ ] VE SDEdit.

- [ ] Sampling from scribble.

- [ ] Image editing only on masked regions.