AIDAS Lab @ SNU - Omni-modal Diffusion

Omni-modal Diffusion Foundation Model

AIDAS Lab @ SNU

Input text

Generated speech

Status

Speech token length

2 512

Total refinement steps

2 512

Block length

2 512

Sampling temperature

0 2

CFG scale

0 6

Gender

Emotion

Speed

Pitch

Examples

Source speech

Reply speech

Status

Reply token length

2 512

Refinement steps

2 512

Block length

2 512

Sampling temperature

0 2

CFG scale

0 6

Examples

Speech input

Transcription

Status

Denoising steps

2 512

Block length

2 512

Max new tokens

2 512

Remasking strategy

Examples

Caption / answer

Status

Denoising steps

2 512

Block length

2 512

Max new tokens

2 512

Examples

Optional instruction

Generated speech

Status

Reply token length

2 512

Refinement steps

2 512

Block length

2 512

Sampling temperature

0 2

CFG scale

0 6

Examples

Prompt

Generated image

Status

Timesteps

4 128

Sampling temperature

0 2

CFG scale

0 8

Examples

Reference image

Editing instruction

Edited image

Status

Timesteps

4 128

Sampling temperature

0 2

CFG scale

0 8

Examples

Reference image	Editing instruction

Source image

Editing instruction

Edited image

Generated text

Status

Text placeholder tokens

8 256

Image timesteps

4 128

Text timesteps

4 128

Sampling temperature

0 2

CFG scale

0 8

Examples

Source image	Editing instruction

Image input

Optional question

Spoken description

Status

Reply token length

2 512

Refinement steps

2 512

Block length

2 512

Sampling temperature

0 2

CFG scale

0 6

Examples

Message

Assistant reply

Status

Reply max tokens

2 512

Refinement steps

2 512

Block length

2 512

Sampling temperature

0 2

Examples

Input image

Question

Answer

Status

Answer max tokens

2 512

Refinement steps

2 512

Block length

2 512

Sampling temperature

0 2

Examples

Input image	Question

Pages:

·

Built with Gradio logo

·

·