You're looking at a specific version of this model. Jump to the model overview.
manuelkugelmann /diffusionlight-turbo:2da0812b
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 |
|---|---|---|---|
| image |
string
|
Input photo. Padded to 1024x1024 if non-square.
|
|
| mode |
None
|
hdr
|
'hdr' → full HDR envmap (.hdr) via 3-bracket merge, for runtime IBL. 'lighting' → darkest-bracket LDR envmap (.png) focused on dominant light sources (windows/lamps), ideal for KB indexing. Only generates the darkest bracket, so ~3× faster than 'hdr'.
|
| envmap_height |
integer
|
256
Min: 64 Max: 512 |
Output equirect height. Width is 2x height.
|
| evs |
string
|
0,-2.5,-5
|
Comma-separated EV brackets. In 'lighting' mode only the most-negative value is used.
|
| denoising_steps |
integer
|
30
Min: 4 Max: 50 |
None
|
| guidance_scale |
number
|
5
Max: 15 |
None
|
| controlnet_scale |
number
|
0.5
Max: 2 |
None
|
| turbo_lora_scale |
number
|
1
Max: 1.5 |
None
|
| exposure_lora_scale |
number
|
0.75
Max: 1.5 |
None
|
| seed |
integer
|
200
|
None
|
| output_ldr_preview |
boolean
|
False
|
In 'hdr' mode, also emit a tonemapped .png alongside the .hdr (ignored in 'lighting' mode — the PNG is the output).
|
Output schema
The shape of the response you’ll get when you run this model with an API.
Schema
{'format': 'uri', 'title': 'Output', 'type': 'string'}