zsxkib/framepack 🔢📝🖼️✓ → 🖼️
About
🕹️FramePack: video diffusion that feels like image diffusion🎥
Example Output
Prompt:
"A woman pets her cat"
Output
Performance Metrics
92.48s
Prediction Time
92.49s
Total Time
All Input Parameters
{
"steps": 25,
"prompt": "A woman pets her cat",
"mp4_crf": 23,
"cfg_scale": 1,
"cfg_rescale": 0,
"input_image": "https://replicate.delivery/pbxt/N0bNhGBlCkyqqTpaGqHXyY9ZXnugIqhzdCO0N6H9VjvPGxrg/image.png",
"use_teacache": true,
"negative_prompt": "",
"latent_window_size": 9,
"distilled_cfg_scale": 10,
"total_video_length_seconds": 3
}
Input Parameters
- seed
- Set for consistent results or leave empty for random
- steps
- More steps give higher quality but take longer
- prompt (required)
- Text description of what you want in the video
- mp4_crf
- Controls video compression - lower values give better quality but larger files
- cfg_scale
- Higher values follow the prompt more closely, lower values are more creative
- cfg_rescale
- Reduces oversaturation at high CFG values
- input_image (required)
- Initial image to start the video from
- use_teacache
- Makes generation faster with minimal quality impact
- negative_prompt
- Things you want to avoid in the video
- latent_window_size
- Controls how video chunks are processed - smaller is faster but may reduce quality
- distilled_cfg_scale
- Controls prompt adherence for distilled model components
- total_video_length_seconds
- How long the video should be in seconds
Output Schema
Output
Example Execution Logs
Using random seed: 2718993960 Encoding text prompts... Processing input image... Encoding initial image with VAE... Encoding image with CLIP Vision model... Starting video generation loop... Processing section 1/2 (padding: 1), is_last_section_logic: False Sampling step 1/25 for section 1 0%| | 0/25 [00:00<?, ?it/s] Sampling step 2/25 for section 1 4%|▍ | 1/25 [00:02<00:49, 2.08s/it] Sampling step 3/25 for section 1 8%|▊ | 2/25 [00:05<01:00, 2.64s/it] Sampling step 4/25 for section 1 12%|█▏ | 3/25 [00:08<01:02, 2.82s/it] Sampling step 5/25 for section 1 16%|█▌ | 4/25 [00:09<00:43, 2.08s/it] Sampling step 6/25 for section 1 20%|██ | 5/25 [00:11<00:41, 2.09s/it] Sampling step 7/25 for section 1 Sampling step 8/25 for section 1 24%|██▍ | 6/25 [00:12<00:32, 1.70s/it] Sampling step 9/25 for section 1 32%|███▏ | 8/25 [00:14<00:23, 1.38s/it] Sampling step 10/25 for section 1 Sampling step 11/25 for section 1 36%|███▌ | 9/25 [00:15<00:20, 1.27s/it] Sampling step 12/25 for section 1 44%|████▍ | 11/25 [00:17<00:16, 1.18s/it] Sampling step 13/25 for section 1 Sampling step 14/25 for section 1 48%|████▊ | 12/25 [00:18<00:14, 1.13s/it] Sampling step 15/25 for section 1 56%|█████▌ | 14/25 [00:20<00:12, 1.10s/it] Sampling step 16/25 for section 1 Sampling step 17/25 for section 1 60%|██████ | 15/25 [00:21<00:10, 1.07s/it] Sampling step 18/25 for section 1 68%|██████▊ | 17/25 [00:23<00:08, 1.06s/it] Sampling step 19/25 for section 1 72%|███████▏ | 18/25 [00:24<00:07, 1.04s/it] Sampling step 20/25 for section 1 76%|███████▌ | 19/25 [00:26<00:07, 1.29s/it] Sampling step 21/25 for section 1 80%|████████ | 20/25 [00:27<00:06, 1.21s/it] Sampling step 22/25 for section 1 84%|████████▍ | 21/25 [00:29<00:05, 1.44s/it] Sampling step 23/25 for section 1 88%|████████▊ | 22/25 [00:32<00:05, 1.88s/it] Sampling step 24/25 for section 1 92%|█████████▏| 23/25 [00:35<00:04, 2.21s/it] Sampling step 25/25 for section 1 96%|█████████▌| 24/25 [00:36<00:01, 1.85s/it] 100%|██████████| 25/25 [00:38<00:00, 1.92s/it] 100%|██████████| 25/25 [00:38<00:00, 1.55s/it] Updated video saved: /tmp/tmp06zetr8l/250514_190911_051_8874_video.mp4, total pixel frames: 33 Processing section 2/2 (padding: 0), is_last_section_logic: True Sampling step 1/25 for section 2 0%| | 0/25 [00:00<?, ?it/s] Sampling step 2/25 for section 2 4%|▍ | 1/25 [00:02<00:50, 2.10s/it] Sampling step 3/25 for section 2 8%|▊ | 2/25 [00:05<01:01, 2.67s/it] Sampling step 4/25 for section 2 12%|█▏ | 3/25 [00:08<01:02, 2.84s/it] Sampling step 5/25 for section 2 16%|█▌ | 4/25 [00:09<00:44, 2.10s/it] Sampling step 6/25 for section 2 20%|██ | 5/25 [00:11<00:42, 2.10s/it] Sampling step 7/25 for section 2 Sampling step 8/25 for section 2 24%|██▍ | 6/25 [00:12<00:32, 1.71s/it] Sampling step 9/25 for section 2 32%|███▏ | 8/25 [00:14<00:23, 1.39s/it] Sampling step 10/25 for section 2 Sampling step 11/25 for section 2 36%|███▌ | 9/25 [00:15<00:20, 1.28s/it] Sampling step 12/25 for section 2 44%|████▍ | 11/25 [00:17<00:16, 1.18s/it] Sampling step 13/25 for section 2 Sampling step 14/25 for section 2 48%|████▊ | 12/25 [00:18<00:14, 1.13s/it] Sampling step 15/25 for section 2 56%|█████▌ | 14/25 [00:20<00:12, 1.10s/it] Sampling step 16/25 for section 2 Sampling step 17/25 for section 2 60%|██████ | 15/25 [00:21<00:10, 1.07s/it] Sampling step 18/25 for section 2 68%|██████▊ | 17/25 [00:23<00:08, 1.06s/it] Sampling step 19/25 for section 2 72%|███████▏ | 18/25 [00:24<00:07, 1.04s/it] Sampling step 20/25 for section 2 76%|███████▌ | 19/25 [00:26<00:07, 1.29s/it] Sampling step 21/25 for section 2 80%|████████ | 20/25 [00:27<00:06, 1.21s/it] Sampling step 22/25 for section 2 84%|████████▍ | 21/25 [00:29<00:05, 1.45s/it] Sampling step 23/25 for section 2 88%|████████▊ | 22/25 [00:32<00:05, 1.89s/it] Sampling step 24/25 for section 2 92%|█████████▏| 23/25 [00:35<00:04, 2.21s/it] Sampling step 25/25 for section 2 96%|█████████▌| 24/25 [00:36<00:01, 1.85s/it] 100%|██████████| 25/25 [00:38<00:00, 1.93s/it] 100%|██████████| 25/25 [00:38<00:00, 1.55s/it] Updated video saved: /tmp/tmp06zetr8l/250514_190911_051_8874_video.mp4, total pixel frames: 73
Version Details
- Version ID
b840152c70e887773e95b24d1f1e8fd2aea448fcf093de801d3627f0a197409f- Version Created
- May 14, 2025