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 |
---|---|---|---|
ckpt |
string
(enum)
|
general_full_v1
Options: face_full_v1, general_full_v1 |
full checkpoint path
|
reload_swinir |
boolean
|
False
|
Flag to reload SwinIR
|
swinir_ckpt |
string
(enum)
|
general_swinir_v1
Options: face_swinir_v1, general_swinir_v1 |
SwinIR checkpoint path
|
input |
string
|
Path to the input image
|
|
steps |
integer
|
50
Min: 1 Max: 100 |
Number of steps
|
sr_scale |
integer
|
4
Min: 1 Max: 4 |
SR scale factor
|
repeat_times |
integer
|
1
Min: 1 Max: 10 |
Number of repetitions
|
disable_preprocess_model |
boolean
|
False
|
Flag to disable model preprocessing i.e. Stage 1 (SwinIR which removes degradations)
|
tiled |
boolean
|
False
|
patch-based sampling
|
tile_size |
integer
|
512
|
TODO
|
tile_stride |
integer
|
256
|
TODO
|
use_guidance |
boolean
|
False
|
Flag to use latent image guidance
|
g_scale |
number
|
0
|
Guidance scale
|
g_t_start |
integer
|
1001
|
Guidance start time
|
g_t_stop |
integer
|
-1
|
Guidance stop time
|
g_space |
string
|
latent
|
Guidance space
|
g_repeat |
integer
|
5
|
Guidance repeat times
|
color_fix_type |
string
(enum)
|
wavelet
Options: wavelet, adain, none |
Type of color correction
|
output |
string
|
.
|
Path to save the output
|
show_lq |
boolean
|
False
|
Flag to show LQ
|
skip_if_exist |
boolean
|
False
|
Flag to skip if output exists
|
seed |
integer
|
231
|
Random seed
|
has_aligned |
boolean
|
False
|
Only applied when checkpoint is set to faces (`ckpt = full_face_v2`) Input are cropped and aligned faces.
|
only_center_face |
boolean
|
False
|
Only applied when checkpoint is set to faces (`ckpt = full_face_v2`) Only restore the center face.
|
detection_model |
string
(enum)
|
retinaface_resnet50
Options: retinaface_resnet50, retinaface_mobile0.25, YOLOv5l, YOLOv5n, dlib |
Only applied when checkpoint is set to faces (`ckpt = full_face_v2`) Face detector model
|
bg_upsampler |
string
(enum)
|
RealESRGAN
Options: DiffBIR, RealESRGAN |
Only applied when checkpoint is set to faces (`ckpt = full_face_v2`) Background upsampler
|
bg_tile |
integer
|
400
|
Only applied when checkpoint is set to faces (`ckpt = full_face_v2`) Tile size for background sampler
|
bg_tile_stride |
integer
|
400
|
Only applied when checkpoint is set to faces (`ckpt = full_face_v2`) Tile stride for background sampler.
|
Output schema
The shape of the response you’ll get when you run this model with an API.
{'items': {'format': 'uri', 'type': 'string'},
'title': 'Output',
'type': 'array'}