lucataco/motif-video 🔢📝✓ → 🖼️

▶️ 41 runs 📅 May 2026 ⚙️ Cog 0.19.1 🔗 GitHub ⚖️ License

text-to-video

Performance

120.8sTypical run time

41Total runs

About

Motif-Video-2B: a 2B-parameter text-to-video diffusion transformer

Example Output

Prompt:

"A vibrant blue jay perches on a slender branch in a sunlit forest, feathers shimmering, soft morning light filtering through leaves"

Output

Performance Metrics

120.80s Prediction Time

217.96s Total Time

All Input Parameters

{
  "prompt": "A vibrant blue jay perches on a slender branch in a sunlit forest, feathers shimmering, soft morning light filtering through leaves",
  "num_frames": 49,
  "num_inference_steps": 30
}

Input Parameters

seed Type: integerDefault: 0: Random seed. Set 0 to randomize.
width Type: integerDefault: 832Range: 320 - 1280: Video width in pixels. 832 is fast 480p test default; 1280 is native 720p.
height Type: integerDefault: 480Range: 320 - 736: Video height in pixels. 480 is fast 480p test default; 736 is native 720p.
prompt (required) Type: string: Text prompt describing the video to generate.
frame_rate Type: integerDefault: 24Range: 8 - 30: Output video frame rate (fps).
num_frames Type: integerDefault: 49Range: 25 - 121: Number of frames. 49 frames @ 24 fps = ~2s test clip; 121 = ~5s full clip.
enable_compile Type: booleanDefault: false
guidance_scale Type: numberDefault: 8Range: 1 - 15: Classifier-free guidance scale.
negative_prompt Type: stringDefault:
num_inference_steps Type: integerDefault: 30Range: 10 - 80: Number of denoising steps. 30 is a good quality/speed tradeoff; bump to 50 for max quality.

Output Schema

Output

Type: string • Format: uri

Example Execution Logs

[predict] seed=3382943503 steps=30 size=832x480 frames=49 compile=False
Guiders are currently an experimental feature under active development. The API is subject to breaking changes in future releases.

  0%|          | 0/30 [00:00<?, ?it/s]Modular Diffusers is currently an experimental feature under active development. The API is subject to breaking changes in future releases.

  3%|▎         | 1/30 [00:03<01:49,  3.76s/it]
  7%|▋         | 2/30 [00:06<01:30,  3.23s/it]
 10%|█         | 3/30 [00:10<01:33,  3.45s/it]
 13%|█▎        | 4/30 [00:14<01:32,  3.56s/it]
 17%|█▋        | 5/30 [00:17<01:30,  3.62s/it]
 20%|██        | 6/30 [00:21<01:27,  3.66s/it]
 23%|██▎       | 7/30 [00:25<01:24,  3.69s/it]
 27%|██▋       | 8/30 [00:29<01:21,  3.71s/it]
 30%|███       | 9/30 [00:32<01:18,  3.73s/it]
 33%|███▎      | 10/30 [00:36<01:14,  3.74s/it]
 37%|███▋      | 11/30 [00:40<01:11,  3.75s/it]
 40%|████      | 12/30 [00:44<01:07,  3.75s/it]
 43%|████▎     | 13/30 [00:47<01:03,  3.76s/it]
 47%|████▋     | 14/30 [00:51<01:00,  3.76s/it]
 50%|█████     | 15/30 [00:55<00:56,  3.77s/it]
 53%|█████▎    | 16/30 [00:59<00:52,  3.77s/it]
 57%|█████▋    | 17/30 [01:02<00:49,  3.78s/it]
 60%|██████    | 18/30 [01:06<00:45,  3.78s/it]
 63%|██████▎   | 19/30 [01:10<00:41,  3.78s/it]
 67%|██████▋   | 20/30 [01:14<00:37,  3.78s/it]
 70%|███████   | 21/30 [01:18<00:34,  3.79s/it]
 73%|███████▎  | 22/30 [01:21<00:30,  3.79s/it]
 77%|███████▋  | 23/30 [01:25<00:26,  3.79s/it]
 80%|████████  | 24/30 [01:29<00:22,  3.79s/it]
 83%|████████▎ | 25/30 [01:33<00:18,  3.79s/it]
 87%|████████▋ | 26/30 [01:37<00:15,  3.79s/it]
 90%|█████████ | 27/30 [01:40<00:11,  3.79s/it]
 93%|█████████▎| 28/30 [01:44<00:07,  3.79s/it]
 97%|█████████▋| 29/30 [01:48<00:03,  3.79s/it]
100%|██████████| 30/30 [01:52<00:00,  3.79s/it]
100%|██████████| 30/30 [01:52<00:00,  3.74s/it]
[predict] generation done in 120.3s

Version Details

Version ID: 5f41c9de463a5a3dc42cd7bec386d24543e4ebebea509e56c9597e53786e539b
Version Created: May 7, 2026

Run on Replicate →