prunaai/vace-1.3b

This is VACE-1.3B model optimised with pruna ai. Wan2.1 VACE is an all-in-one model for video creation and editing.

Capabilities

Seed

Community model (estimated from hardware time)

Name	Type	Description	Default	Constraints
`prompt`*	string	Prompt	`—`	—
`frame_num`	integer	Number of frames to generate.	`81`	—
`sample_guide_scale`	number	Sample guide scale	`5`	—
`sample_shift`	integer	Sample shift	`16`	—
`sample_solver`	string	Sample solver	`"unipc"`	unipcdpm++
`sample_steps`	integer	Sample steps	`50`	—
`seed`	integer	Random seed (-1 for random)	`-1`	—
`size`	string	Output resolution	`"480*832"`	480832832480
`speed_mode`	string	Speed optimization level	`"Lightly Juiced 🍊 (more consistent)"`	Lightly Juiced 🍊 (more consistent)Juiced 🔥 (more speed)Extra Juiced 🚀 (even more speed)
`src_mask`	string(uri)	Input mask video to edit.	`—`	—
`src_ref_images`	array	Input reference images to edit.	`—`	—
`src_video`	string(uri)	Input video to edit.	`—`	—

promptrequiredstring

Prompt

frame_numinteger

Number of frames to generate.

Default: 81

sample_guide_scalenumber

Sample guide scale

Default: 5

sample_shiftinteger

Sample shift

Default: 16

sample_solverstring

Sample solver

Default: "unipc"

unipcdpm++

sample_stepsinteger

Sample steps

Default: 50

seedinteger

Random seed (-1 for random)

Default: -1

sizestring

Output resolution

Default: "480*832"

480*832832*480

speed_modestring

Speed optimization level

Default: "Lightly Juiced 🍊 (more consistent)"

Lightly Juiced 🍊 (more consistent)Juiced 🔥 (more speed)Extra Juiced 🚀 (even more speed)

src_maskstring

Input mask video to edit.

src_ref_imagesarray

Input reference images to edit.

src_videostring

Input video to edit.

Version: ecf7f253f49cUpdated: 8/1/2026731 runs