nodes/hao-ai-lab-FastVideo
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[feat] add support for Kandinsky 5 Pro, Lite I2V

Browse Source
This commit is contained in:
leffff
2026-07-05 14:58:12 -07:00
committed by SolitaryThinker
parent aa893056e8
commit 562b951208
4 changed files with 190 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files
examples/inference/basic/basic_kandinsky5_i2v.py
+7 -8
View File
@@ -1,13 +1,15 @@
from fastvideo import VideoGenerator
OUTPUT_PATH = "video_samples_kandinsky5_2B_i2v"
OUTPUT_PATH = "video_samples_kandinsky5_i2v"
IMAGE_PATH = "assets/girl.png"
def main():
generator = VideoGenerator.from_pretrained(
"kandinskylab/Kandinsky-5.0-I2V-Lite-sft-5s-Diffusers",
"kandinskylab/Kandinsky-5.0-I2V-Pro-distilled-5s-Diffusers",
# "kandinskylab/Kandinsky-5.0-I2V-Pro-sft-5s-Diffusers"
# "kandinskylab/Kandinsky-5.0-I2V-Lite-5s-Diffusers"
num_gpus=1,
use_fsdp_inference=False,
dit_cpu_offload=False,
@@ -18,18 +20,15 @@ def main():
)
prompt = (
"A woman sits at a wooden table by the window in a cozy café. She reaches out "
"with her right hand, picks up the white coffee cup from the saucer, and gently "
"brings it to her lips to take a sip. After drinking, she places the cup back on "
"the table and looks out the window, enjoying the peaceful atmosphere."
"A woman stands up and walks away"
)
_ = generator.generate_video(
prompt,
image_path=IMAGE_PATH,
output_path=OUTPUT_PATH,
save_video=True,
height=512,
width=768,
height=1024,
width=1024,
num_frames=121,
)
examples/inference/basic/basic_kandinsky5_t2v.py
+5 -2
View File
@@ -1,10 +1,13 @@
from fastvideo import VideoGenerator
OUTPUT_PATH = "video_samples_kandinsky5_2B_t2v"
OUTPUT_PATH = "video_samples_kandinsky5_t2v"
def main():
generator = VideoGenerator.from_pretrained(
"kandinskylab/Kandinsky-5.0-T2V-Lite-sft-5s-Diffusers",
"kandinskylab/Kandinsky-5.0-T2V-Lite-sft-5s-Diffusers",
# "kandinskylab/Kandinsky-5.0-T2V-Pro-sft-5s-Diffusers"
# "kandinskylab/Kandinsky-5.0-T2V-Lite-Distilled-5s-Diffusers"
# "kandinskylab/Kandinsky-5.0-T2V-Pro-Distilled-5s-Diffusers"
num_gpus=1,
use_fsdp_inference=False,
dit_cpu_offload=False,
fastvideo/pipelines/basic/kandinsky5/presets.py
+91 -1
View File
@@ -34,6 +34,60 @@ KANDINSKY5_T2V_LITE_5S = InferencePreset(
},
)
KANDINSKY5_T2V_LITE_DISTILLED_5S = InferencePreset(
name="kandinsky5_t2v_lite_distilled_5s",
version=1,
model_family="kandinsky5",
description="Kandinsky-5.0 Lite T2V Distilled 5s",
workload_type="t2v",
stage_schemas=(_DENOISE_STAGE, ),
defaults={
"height": 512,
"width": 768,
"num_frames": 121,
"fps": 24,
"guidance_scale": 1.0,
"num_inference_steps": 16,
"negative_prompt": _NEGATIVE_PROMPT,
},
)
KANDINSKY5_T2V_PRO_5S = InferencePreset(
name="kandinsky5_t2v_pro_5s",
version=1,
model_family="kandinsky5",
description="Kandinsky-5.0 Pro T2V 5s",
workload_type="t2v",
stage_schemas=(_DENOISE_STAGE, ),
defaults={
"height": 512,
"width": 768,
"num_frames": 121,
"fps": 24,
"guidance_scale": 5.0,
"num_inference_steps": 50,
"negative_prompt": _NEGATIVE_PROMPT,
},
)
KANDINSKY5_T2V_PRO_DISTILLED_5S = InferencePreset(
name="kandinsky5_t2v_pro_distilled_5s",
version=1,
model_family="kandinsky5",
description="Kandinsky-5.0 Pro T2V Distilled 5s",
workload_type="t2v",
stage_schemas=(_DENOISE_STAGE, ),
defaults={
"height": 512,
"width": 768,
"num_frames": 121,
"fps": 24,
"guidance_scale": 1.0,
"num_inference_steps": 16,
"negative_prompt": _NEGATIVE_PROMPT,
},
)
KANDINSKY5_I2V_LITE_5S = InferencePreset(
name="kandinsky5_i2v_lite_5s",
version=1,
@@ -52,4 +106,40 @@ KANDINSKY5_I2V_LITE_5S = InferencePreset(
},
)
ALL_PRESETS = (KANDINSKY5_T2V_LITE_5S, KANDINSKY5_I2V_LITE_5S)
KANDINSKY5_I2V_PRO_5S = InferencePreset(
name="kandinsky5_i2v_pro_5s",
version=1,
model_family="kandinsky5",
description="Kandinsky-5.0 Pro I2V 5s",
workload_type="i2v",
stage_schemas=(_DENOISE_STAGE, ),
defaults={
"height": 512,
"width": 768,
"num_frames": 121,
"fps": 24,
"guidance_scale": 5.0,
"num_inference_steps": 50,
"negative_prompt": _NEGATIVE_PROMPT,
},
)
KANDINSKY5_I2V_PRO_DISTILLED_5S = InferencePreset(
name="kandinsky5_i2v_pro_distilled_5s",
version=1,
model_family="kandinsky5",
description="Kandinsky-5.0 Pro I2V Distilled 5s",
workload_type="i2v",
stage_schemas=(_DENOISE_STAGE, ),
defaults={
"height": 512,
"width": 768,
"num_frames": 121,
"fps": 24,
"guidance_scale": 1.0,
"num_inference_steps": 16,
"negative_prompt": _NEGATIVE_PROMPT,
},
)
ALL_PRESETS = (KANDINSKY5_T2V_LITE_5S, KANDINSKY5_T2V_LITE_DISTILLED_5S, KANDINSKY5_T2V_PRO_5S, KANDINSKY5_T2V_PRO_DISTILLED_5S, KANDINSKY5_I2V_LITE_5S, KANDINSKY5_I2V_PRO_5S, KANDINSKY5_I2V_PRO_DISTILLED_5S)
fastvideo/registry.py
+87 -3
View File
@@ -504,6 +504,7 @@ def _register_configs() -> None:
workload_types=(WorkloadType.T2V, ),
hf_model_paths=[
"kandinskylab/Kandinsky-5.0-T2V-Lite-sft-5s-Diffusers",
],
model_detectors=[
lambda path: any(token in path.lower() for token in ("kandinsky5", "kandinsky-5"))
@@ -514,14 +515,65 @@ def _register_configs() -> None:
pipeline_cls_name="Kandinsky5T2VPipeline",
)
# Kandinsky5 I2V
# Kandinsky5 Pro T2V
register_configs(
sampling_param_cls=None,
pipeline_config_cls=Kandinsky5T2VConfig,
workload_types=(WorkloadType.T2V, ),
hf_model_paths=[
"kandinskylab/Kandinsky-5.0-T2V-Pro-sft-5s-Diffusers",
],
model_detectors=[
lambda path: any(token in path.lower() for token in ("kandinsky5", "kandinsky-5"))
and "i2v" not in path.lower(),
],
model_family="kandinsky5",
default_preset="kandinsky5_t2v_pro_5s",
)
# Kandinsky5 Lite T2V Distilled
register_configs(
sampling_param_cls=None,
pipeline_config_cls=Kandinsky5T2VConfig,
workload_types=(WorkloadType.T2V, ),
hf_model_paths=[
"kandinskylab/Kandinsky-5.0-T2V-Lite-distilled-5s-Diffusers",
],
model_detectors=[
lambda path: any(token in path.lower() for token in ("kandinsky5", "kandinsky-5"))
and "i2v" not in path.lower(),
],
model_family="kandinsky5",
default_preset="kandinsky5_t2v_lite_distilled_5s",
)
# Kandinsky5 Pro T2V Distilled
register_configs(
sampling_param_cls=None,
pipeline_config_cls=Kandinsky5T2VConfig,
workload_types=(WorkloadType.T2V, ),
hf_model_paths=[
"kandinskylab/Kandinsky-5.0-T2V-Pro-distilled-5s-Diffusers",
],
model_detectors=[
lambda path: any(token in path.lower() for token in ("kandinsky5", "kandinsky-5"))
and "i2v" not in path.lower(),
],
model_family="kandinsky5",
default_preset="kandinsky5_t2v_pro_distilled_5s",
)
# Kandinsky5 Lite I2V
register_configs(
sampling_param_cls=None,
pipeline_config_cls=Kandinsky5I2VConfig,
workload_types=(WorkloadType.I2V, ),
hf_model_paths=[
"kandinskylab/Kandinsky-5.0-I2V-Lite-sft-5s-Diffusers",
"kandinskylab/Kandinsky-5.0-I2V-Pro-sft-5s-Diffusers",
"kandinskylab/Kandinsky-5.0-I2V-Lite-5s-Diffusers"
],
model_detectors=[
lambda path: any(token in path.lower() for token in ("kandinsky5", "kandinsky-5"))
@@ -532,6 +584,38 @@ def _register_configs() -> None:
pipeline_cls_name="Kandinsky5I2VPipeline",
)
# Kandinsky5 Pro I2V
register_configs(
sampling_param_cls=None,
pipeline_config_cls=Kandinsky5I2VConfig,
workload_types=(WorkloadType.I2V, ),
hf_model_paths=[
"kandinskylab/Kandinsky-5.0-I2V-Pro-sft-5s-Diffusers"
],
model_detectors=[
lambda path: any(token in path.lower() for token in ("kandinsky5", "kandinsky-5"))
and "i2v" in path.lower(),
],
model_family="kandinsky5",
default_preset="kandinsky5_i2v_pro_5s",
)
# Kandinsky5 Pro I2V Distilled
register_configs(
sampling_param_cls=None,
pipeline_config_cls=Kandinsky5I2VConfig,
workload_types=(WorkloadType.I2V, ),
hf_model_paths=[
"kandinskylab/Kandinsky-5.0-I2V-Pro-distilled-5s-Diffusers"
],
model_detectors=[
lambda path: any(token in path.lower() for token in ("kandinsky5", "kandinsky-5"))
and "i2v" in path.lower(),
],
model_family="kandinsky5",
default_preset="kandinsky5_i2v_pro_distilled_5s",
)
# LongCat (T2V, I2V, VC use same config; workload varies by path)
register_configs(
sampling_param_cls=None,