Skip to content

The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.

[Partner Nodes] feat(Gemini): add Omni Flash 1.2 to the Omni video node (#17560)

Alexander Piskun6b2031647499
diff --git a/comfy_api_nodes/nodes_gemini.py b/comfy_api_nodes/nodes_gemini.py
index 1862fb35c..e3850af42 100644
--- a/comfy_api_nodes/nodes_gemini.py
+++ b/comfy_api_nodes/nodes_gemini.py
@@ -1819,6 +1819,7 @@ OMNI_URI_DELIVERY_RESOLUTIONS = ("1080p", "4k")
OMNI_MODELS: dict[str, str] = {
"Omni Flash": "gemini-omni-flash-preview",
"Omni Flash 1.1": "gemini-omni-1.1-flash",
+ "Omni Flash 1.2": "gemini-omni-1.2-flash",
}
@@ -2106,6 +2107,56 @@ def _omni_v2_flash_1_1_inputs() -> list[Input]:
]
+def _omni_v2_flash_1_2_inputs() -> list[Input]:
+ return [
+ IO.String.Input(
+ "prompt",
+ multiline=True,
+ default="",
+ tooltip="Describe the video to generate, or the edit to apply to an attached video. Specify the "
+ 'length directly in the prompt, e.g. "a 6-second clip" or, for the \'extend\' task, "extend by 5 seconds"; '
+ "the generated length may be 3-15 seconds, and 'extend' adds at most 10 seconds.",
+ ),
+ IO.Combo.Input(
+ "resolution",
+ options=["360p", "720p", "1080p", "4k"],
+ default="720p",
+ tooltip="Output resolution.",
+ ),
+ IO.Combo.Input(
+ "aspect_ratio",
+ options=["16:9", "9:16"],
+ default="16:9",
+ tooltip="Output aspect ratio: 16:9 (landscape) or 9:16 (portrait). "
+ "The 'edit' and 'extend' tasks keep the aspect ratio of the input video instead.",
+ ),
+ _omni_task_input(with_extend=True),
+ IO.Autogrow.Input(
+ "images",
+ template=IO.Autogrow.TemplateNames(
+ IO.Image.Input("image"),
+ names=[f"image_{i}" for i in range(1, OMNI_MAX_IMAGES + 1)],
+ min=0,
+ ),
+ tooltip=f"Optional reference image(s) to guide or animate the video. Up to {OMNI_MAX_IMAGES} images; "
+ "with the 'image_to_video' task the first one is the starting frame and an optional second one "
+ "is the ending frame.",
+ ),
+ IO.Autogrow.Input(
+ "videos",
+ template=IO.Autogrow.TemplateNames(
+ IO.Video.Input("video"),
+ names=[f"video_{i}" for i in range(1, OMNI_MAX_VIDEOS + 1)],
+ min=0,
+ ),
+ tooltip=f"Optional reference video(s) to guide or edit. Up to {OMNI_MAX_VIDEOS} videos, "
+ "each up to 15 seconds long. The 'extend' task accepts up to 25 seconds, "
+ "and its result can be at most 30 seconds.",
+ ),
+ _omni_seed_input(),
+ ]
+
+
class GeminiVideoOmniV2(IO.ComfyNode):
@@ -2118,11 +2169,12 @@ class GeminiVideoOmniV2(IO.ComfyNode):
essentials_category="Video Generation",
description="Generate a video with audio from a text prompt using Google's Gemini Omni Flash models. "
"Optionally provide reference images and/or videos to guide or edit the result. Describe the desired "
- "length (3-10s) directly in the prompt.",
+ "length directly in the prompt (3-15s for Omni Flash 1.2, 3-10s for older models).",
inputs=[
IO.DynamicCombo.Input(
"model",
options=[
+ IO.DynamicCombo.Option("Omni Flash 1.2", _omni_v2_flash_1_2_inputs()),
IO.DynamicCombo.Option("Omni Flash 1.1", _omni_v2_flash_1_1_inputs()),
IO.DynamicCombo.Option("Omni Flash", _omni_v2_flash_inputs()),
],
@@ -2143,9 +2195,11 @@ class GeminiVideoOmniV2(IO.ComfyNode):
depends_on=IO.PriceBadgeDepends(widgets=["model", "model.resolution"]),
expr="""
(
- $prices := {"360p": 0.0483, "720p": 0.1449, "1080p": 0.2174, "4k": 0.4349};
+ $prices := widgets.model = "omni flash 1.2"
+ ? {"360p": 0.0331, "720p": 0.0994, "1080p": 0.1491, "4k": 0.2982}
+ : {"360p": 0.0483, "720p": 0.1449, "1080p": 0.2174, "4k": 0.4349};
$r := $lookup(widgets, "model.resolution");
- {"type":"usd","usd": $r ? $lookup($prices, $r) : 0.1449,
+ {"type":"usd","usd": $lookup($prices, $r ? $r : "720p"),
"format":{"suffix":"/second","approximate":true}}
)
""",
@@ -2166,8 +2220,11 @@ class GeminiVideoOmniV2(IO.ComfyNode):
raise ValueError(f"The current maximum number of supported images is {OMNI_MAX_IMAGES}.")
if len(videos) > OMNI_MAX_VIDEOS:
raise ValueError(f"The current maximum number of supported videos is {OMNI_MAX_VIDEOS}.")
+ max_video_duration = 10.1
+ if model["model"] == "Omni Flash 1.2":
+ max_video_duration = 25.1 if task == "extend" else 15.1
for video in videos:
- validate_video_duration(video, max_duration=10.1)
+ validate_video_duration(video, max_duration=max_video_duration)
if task == "text_to_video" and (images or videos):
raise ValueError("The 'text_to_video' task generates from the prompt alone; detach the reference media.")
if task == "image_to_video" and videos: