You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We are missing a small API sample. I used the recorder to figure out what it is sending out and expecting, but it seems to be something else hidden. I checked all files in this API requests and they are fully alpha transparent images. None make sense and no idea what they are used for since they are none of the images (source and target) I placed.
An example on how to do a basic face swap using the API would be great appreciated.
Here is the API that the recorder returned. They contain an imagemask with 3 files, but those files are transparent and who knows what they are...
It sounds like you want to integrate this into something bigger, you know that this can't be used for commercial stuff?
Anyway, to remotely control this you'd need to do the same steps as the user, meaning you don't have just 1 API call but several.
Never used this myself but roughly (for a minimal example without selecting faces whatsoever) it should be calling:
We are missing a small API sample. I used the recorder to figure out what it is sending out and expecting, but it seems to be something else hidden. I checked all files in this API requests and they are fully alpha transparent images. None make sense and no idea what they are used for since they are none of the images (source and target) I placed.
An example on how to do a basic face swap using the API would be great appreciated.
Here is the API that the recorder returned. They contain an imagemask with 3 files, but those files are transparent and who knows what they are...
client.predict(
enhancer="GFPGAN",
detection="First found",
keep_frames=True,
wait_after_extraction=False,
skip_audio=True,
face_distance=1,
blend_ratio=1,
selected_mask_engine="None",
clip_text="cup,hands,hair,banana",
processing_method="In-Memory processing",
no_face_action="Use untouched original frame",
vr_mode=False,
autorotate=True,
num_swap_steps=1,
imagemask={"background":file('http://127.0.0.1:7862/file=D:\\roop-unleashed\\roop-unleashed\\installer\\roop-unleashed\\temp\\d0b0e0d125e4fdaed3846375fae0b0c75b354b6b\\background.png'),"layers":[file('http://127.0.0.1:7862/file=D:\\roop-unleashed\\roop-unleashed\\installer\\roop-unleashed\\temp\\d0b0e0d125e4fdaed3846375fae0b0c75b354b6b\\layer_0.png')],"composite":file('http://127.0.0.1:7862/file=D:\\roop-unleashed\\roop-unleashed\\installer\\roop-unleashed\\temp\\d0b0e0d125e4fdaed3846375fae0b0c75b354b6b\\composite.png')},
api_name="/start_swap"
)
client.predict(
files=[file('http://127.0.0.1:7862/file=D:\roop-unleashed\roop-unleashed\installer\roop-unleashed\temp\714acf6b840fb080b0843c04cec3662112836674\65e2163827107599e688_bf2c2889db294ac6ad0c1a07bfab0b20_23-28-20.webp')],
api_name="/on_resultfiles_finished"
)
The text was updated successfully, but these errors were encountered: