You're looking at a specific version of this model. Jump to the model overview.
Input schema
The fields you can use to run this model with an API. If you don’t give a value for a field its default value will be used.
| Field | Type | Default value | Description |
|---|---|---|---|
| input_image |
string
|
Initial image for video generation.
|
|
| prompt |
string
|
Text prompt describing the desired video content.
|
|
| negative_prompt |
string
|
|
Text prompt specifying elements to avoid in the video.
|
| seed |
integer
|
Random seed for reproducible results. Leave blank for a random seed.
|
|
| total_video_length_seconds |
number
|
3
Min: 1 Max: 10 |
Total video length in seconds. Max 10s is recommended for API stability and performance.
|
| latent_window_size |
integer
|
9
Min: 1 Max: 16 |
Size of latent window (video chunk) for progressive generation. Default (9) is tuned for the model.
|
| steps |
integer
|
25
Min: 1 Max: 50 |
Number of diffusion steps per video segment. More steps can increase detail but take longer; fewer steps are faster but may reduce quality.
|
| cfg_scale |
number
|
1
Min: 1 Max: 32 |
Classifier-Free Guidance scale. Higher values enforce stronger prompt adherence; lower values encourage more creativity.
|
| distilled_cfg_scale |
number
|
10
Min: 1 Max: 32 |
Distilled CFG scale. Similar to `cfg_scale`, influences prompt adherence for the model's distilled components.
|
| cfg_rescale |
number
|
0
Max: 1 |
Guidance rescale factor (0.0-1.0). Helps prevent oversaturation at high CFG values; 0.0 to disable.
|
| use_teacache |
boolean
|
True
|
Enable TeaCache for potentially faster inference (may slightly alter results).
|
| mp4_crf |
integer
|
23
Max: 51 |
MP4 Constant Rate Factor (0-51). Lower values mean better video quality and larger file size; ~23 is a good balance.
|
Output schema
The shape of the response you’ll get when you run this model with an API.
{'format': 'uri', 'title': 'Output', 'type': 'string'}