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 |
---|---|---|---|
gen_model |
string
(enum)
|
icgan
Options: icgan, cc_icgan |
Select type of IC-GAN model. "icgan" is conditioned on an instance; "cc_icgan" is conditioned on both instance and a class index
|
image |
string
|
Input image Instance. Optional. Decide which instance to condition on. Leave it empty for condition on input feature instance (below)
|
|
input_feature_index |
integer
|
0
|
Input Feature Instance. The is valid when input image is not given.Choose an integer from 0 to 999. This will change the instance conditioning and therefore the style and semantics of the generated images. This will select one of the 1000 instance features pre-selected from ImageNet using k-means.
|
class_index |
integer
|
0
|
only valid for gen_model="cc_icgan". Choose an integer from 0 to 999. This will change the ImageNet class to condition on. Refer to this link (https://gist.github.com/yrevar/942d3a0ac09ec9e5eb3a) for corresponding class names and indexes
|
num_samples_total |
integer
|
160
|
total generated samples
|
num_samples_ranked |
integer
|
16
|
the number of generated images to output in a mosaic. These are the ones that scored a higher cosine similarity with the conditioning instance, out of num_samples_total generated samples. A ratio of 10:1 (num_samples_total:num_samples_ranked) is recommended
|
truncation |
number
|
0.7
|
choose from 0 to 1 to apply the truncation trick. truncation=1 will provide more diverse but possibly poorer quality images. Truncation values between 0.7 and 0.9 seem to empirically work well
|
seed |
integer
|
0
|
seed=0 means no seed
|
Output schema
The shape of the response you’ll get when you run this model with an API.
Schema
{'items': {'properties': {'file': {'format': 'uri',
'type': 'string',
'x-order': 0},
'text': {'type': 'string', 'x-order': 1}},
'type': 'object'},
'type': 'array',
'x-cog-array-type': 'iterator'}