adirik/masactrl-stable-diffusion-v1-4 🔢🖼️📝 → 🖼️
About
Edit real or generated images
Example Output
Output
Performance Metrics
25.26s
Prediction Time
25.30s
Total Time
All Input Parameters
{
"source_image": "https://replicate.delivery/pbxt/JzivdO6ZRyGFooLwygTxvFlaTjctken1O6FxdIbLQGVZFhPN/corgi.jpg",
"target_prompt": "a photo of a running corgi",
"guidance_scale": 7.5,
"masactrl_start_step": 4,
"num_inference_steps": 50,
"masactrl_start_layer": 10
}
Input Parameters
- seed
- Random seed.A random seed will be used if not provided.
- source_image
- Image to edit for image editing mode.
- source_prompt
- Prompt for the first image for consistent image synthesis mode
- target_prompt (required)
- If consistent image syhnthesis mode, prompt for the second image. If image editing mode, prompt for the target image.
- guidance_scale
- Scale for classifier-free guidance
- masactrl_start_step
- The step to start mutual self-attention control. It should be lower than num_inference_steps
- num_inference_steps
- Number of denoising steps
- masactrl_start_layer
- The layer to start mutual self-attention control
Output Schema
Output
Example Execution Logs
Seed set to 2283628953 input text embeddings : torch.Size([1, 77, 768]) latents shape: torch.Size([1, 4, 64, 64]) Valid timesteps: tensor([ 1, 21, 41, 61, 81, 101, 121, 141, 161, 181, 201, 221, 241, 261, 281, 301, 321, 341, 361, 381, 401, 421, 441, 461, 481, 501, 521, 541, 561, 581, 601, 621, 641, 661, 681, 701, 721, 741, 761, 781, 801, 821, 841, 861, 881, 901, 921, 941, 961, 981]) DDIM Inversion: 0%| | 0/50 [00:00<?, ?it/s] DDIM Inversion: 4%|▍ | 2/50 [00:00<00:04, 10.60it/s] DDIM Inversion: 8%|▊ | 4/50 [00:00<00:05, 8.22it/s] DDIM Inversion: 10%|█ | 5/50 [00:00<00:05, 7.85it/s] DDIM Inversion: 12%|█▏ | 6/50 [00:00<00:05, 7.61it/s] DDIM Inversion: 14%|█▍ | 7/50 [00:00<00:05, 7.44it/s] DDIM Inversion: 16%|█▌ | 8/50 [00:01<00:05, 7.32it/s] DDIM Inversion: 18%|█▊ | 9/50 [00:01<00:05, 7.25it/s] DDIM Inversion: 20%|██ | 10/50 [00:01<00:05, 7.20it/s] DDIM Inversion: 22%|██▏ | 11/50 [00:01<00:05, 7.15it/s] DDIM Inversion: 24%|██▍ | 12/50 [00:01<00:05, 7.13it/s] DDIM Inversion: 26%|██▌ | 13/50 [00:01<00:05, 7.10it/s] DDIM Inversion: 28%|██▊ | 14/50 [00:01<00:05, 7.08it/s] DDIM Inversion: 30%|███ | 15/50 [00:02<00:04, 7.08it/s] DDIM Inversion: 32%|███▏ | 16/50 [00:02<00:04, 7.07it/s] DDIM Inversion: 34%|███▍ | 17/50 [00:02<00:04, 7.07it/s] DDIM Inversion: 36%|███▌ | 18/50 [00:02<00:04, 7.07it/s] DDIM Inversion: 38%|███▊ | 19/50 [00:02<00:04, 7.07it/s] DDIM Inversion: 40%|████ | 20/50 [00:02<00:04, 7.07it/s] DDIM Inversion: 42%|████▏ | 21/50 [00:02<00:04, 7.07it/s] DDIM Inversion: 44%|████▍ | 22/50 [00:03<00:03, 7.06it/s] DDIM Inversion: 46%|████▌ | 23/50 [00:03<00:03, 7.06it/s] DDIM Inversion: 48%|████▊ | 24/50 [00:03<00:03, 7.06it/s] DDIM Inversion: 50%|█████ | 25/50 [00:03<00:03, 7.06it/s] DDIM Inversion: 52%|█████▏ | 26/50 [00:03<00:03, 7.06it/s] DDIM Inversion: 54%|█████▍ | 27/50 [00:03<00:03, 7.06it/s] DDIM Inversion: 56%|█████▌ | 28/50 [00:03<00:03, 7.06it/s] DDIM Inversion: 58%|█████▊ | 29/50 [00:04<00:02, 7.06it/s] DDIM Inversion: 60%|██████ | 30/50 [00:04<00:02, 7.07it/s] DDIM Inversion: 62%|██████▏ | 31/50 [00:04<00:02, 7.06it/s] DDIM Inversion: 64%|██████▍ | 32/50 [00:04<00:02, 7.05it/s] DDIM Inversion: 66%|██████▌ | 33/50 [00:04<00:02, 7.06it/s] DDIM Inversion: 68%|██████▊ | 34/50 [00:04<00:02, 7.05it/s] DDIM Inversion: 70%|███████ | 35/50 [00:04<00:02, 7.05it/s] DDIM Inversion: 72%|███████▏ | 36/50 [00:05<00:01, 7.05it/s] DDIM Inversion: 74%|███████▍ | 37/50 [00:05<00:01, 7.05it/s] DDIM Inversion: 76%|███████▌ | 38/50 [00:05<00:01, 7.05it/s] DDIM Inversion: 78%|███████▊ | 39/50 [00:05<00:01, 7.06it/s] DDIM Inversion: 80%|████████ | 40/50 [00:05<00:01, 7.06it/s] DDIM Inversion: 82%|████████▏ | 41/50 [00:05<00:01, 7.07it/s] DDIM Inversion: 84%|████████▍ | 42/50 [00:05<00:01, 7.07it/s] DDIM Inversion: 86%|████████▌ | 43/50 [00:05<00:00, 7.07it/s] DDIM Inversion: 88%|████████▊ | 44/50 [00:06<00:00, 7.06it/s] DDIM Inversion: 90%|█████████ | 45/50 [00:06<00:00, 7.06it/s] DDIM Inversion: 92%|█████████▏| 46/50 [00:06<00:00, 7.05it/s] DDIM Inversion: 94%|█████████▍| 47/50 [00:06<00:00, 7.04it/s] DDIM Inversion: 96%|█████████▌| 48/50 [00:06<00:00, 7.04it/s] DDIM Inversion: 98%|█████████▊| 49/50 [00:06<00:00, 7.04it/s] DDIM Inversion: 100%|██████████| 50/50 [00:06<00:00, 7.05it/s] DDIM Inversion: 100%|██████████| 50/50 [00:06<00:00, 7.16it/s] MasaCtrl at denoising steps: [4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49] MasaCtrl at U-Net layers: [10, 11, 12, 13, 14, 15] input text embeddings : torch.Size([2, 77, 768]) latents shape: torch.Size([2, 4, 64, 64]) DDIM Sampler: 0%| | 0/50 [00:00<?, ?it/s] DDIM Sampler: 4%|▍ | 2/50 [00:00<00:08, 5.63it/s] DDIM Sampler: 6%|▌ | 3/50 [00:00<00:10, 4.66it/s] DDIM Sampler: 8%|▊ | 4/50 [00:00<00:10, 4.27it/s] DDIM Sampler: 10%|█ | 5/50 [00:01<00:11, 4.04it/s] DDIM Sampler: 12%|█▏ | 6/50 [00:01<00:12, 3.64it/s] DDIM Sampler: 14%|█▍ | 7/50 [00:01<00:12, 3.43it/s] DDIM Sampler: 16%|█▌ | 8/50 [00:02<00:12, 3.30it/s] DDIM Sampler: 18%|█▊ | 9/50 [00:02<00:12, 3.21it/s] DDIM Sampler: 20%|██ | 10/50 [00:02<00:12, 3.16it/s] DDIM Sampler: 22%|██▏ | 11/50 [00:03<00:12, 3.12it/s] DDIM Sampler: 24%|██▍ | 12/50 [00:03<00:12, 3.10it/s] DDIM Sampler: 26%|██▌ | 13/50 [00:03<00:12, 3.08it/s] DDIM Sampler: 28%|██▊ | 14/50 [00:04<00:11, 3.07it/s] DDIM Sampler: 30%|███ | 15/50 [00:04<00:11, 3.06it/s] DDIM Sampler: 32%|███▏ | 16/50 [00:04<00:11, 3.06it/s] DDIM Sampler: 34%|███▍ | 17/50 [00:05<00:10, 3.05it/s] DDIM Sampler: 36%|███▌ | 18/50 [00:05<00:10, 3.05it/s] DDIM Sampler: 38%|███▊ | 19/50 [00:05<00:10, 3.05it/s] DDIM Sampler: 40%|████ | 20/50 [00:06<00:09, 3.05it/s] DDIM Sampler: 42%|████▏ | 21/50 [00:06<00:09, 3.05it/s] DDIM Sampler: 44%|████▍ | 22/50 [00:06<00:09, 3.05it/s] DDIM Sampler: 46%|████▌ | 23/50 [00:07<00:08, 3.05it/s] DDIM Sampler: 48%|████▊ | 24/50 [00:07<00:08, 3.04it/s] DDIM Sampler: 50%|█████ | 25/50 [00:07<00:08, 3.04it/s] DDIM Sampler: 52%|█████▏ | 26/50 [00:08<00:07, 3.04it/s] DDIM Sampler: 54%|█████▍ | 27/50 [00:08<00:07, 3.04it/s] DDIM Sampler: 56%|█████▌ | 28/50 [00:08<00:07, 3.04it/s] DDIM Sampler: 58%|█████▊ | 29/50 [00:09<00:06, 3.04it/s] DDIM Sampler: 60%|██████ | 30/50 [00:09<00:06, 3.04it/s] DDIM Sampler: 62%|██████▏ | 31/50 [00:09<00:06, 3.04it/s] DDIM Sampler: 64%|██████▍ | 32/50 [00:10<00:05, 3.04it/s] DDIM Sampler: 66%|██████▌ | 33/50 [00:10<00:05, 3.04it/s] DDIM Sampler: 68%|██████▊ | 34/50 [00:10<00:05, 3.04it/s] DDIM Sampler: 70%|███████ | 35/50 [00:11<00:04, 3.04it/s] DDIM Sampler: 72%|███████▏ | 36/50 [00:11<00:04, 3.04it/s] DDIM Sampler: 74%|███████▍ | 37/50 [00:11<00:04, 3.04it/s] DDIM Sampler: 76%|███████▌ | 38/50 [00:12<00:03, 3.04it/s] DDIM Sampler: 78%|███████▊ | 39/50 [00:12<00:03, 3.04it/s] DDIM Sampler: 80%|████████ | 40/50 [00:12<00:03, 3.04it/s] DDIM Sampler: 82%|████████▏ | 41/50 [00:12<00:02, 3.04it/s] DDIM Sampler: 84%|████████▍ | 42/50 [00:13<00:02, 3.04it/s] DDIM Sampler: 86%|████████▌ | 43/50 [00:13<00:02, 3.04it/s] DDIM Sampler: 88%|████████▊ | 44/50 [00:13<00:01, 3.04it/s] DDIM Sampler: 90%|█████████ | 45/50 [00:14<00:01, 3.04it/s] DDIM Sampler: 92%|█████████▏| 46/50 [00:14<00:01, 3.04it/s] DDIM Sampler: 94%|█████████▍| 47/50 [00:14<00:00, 3.04it/s] DDIM Sampler: 96%|█████████▌| 48/50 [00:15<00:00, 3.04it/s] DDIM Sampler: 98%|█████████▊| 49/50 [00:15<00:00, 3.04it/s] DDIM Sampler: 100%|██████████| 50/50 [00:15<00:00, 3.04it/s] DDIM Sampler: 100%|██████████| 50/50 [00:15<00:00, 3.13it/s]
Version Details
- Version ID
4e86d80ab64a8395e7fd327d34fe85d240a3d9e8706b7144864ba981eba3dfa6- Version Created
- December 5, 2023