Huanyan-Studio / demo_app.py
Sergidev's picture
v2 beta
341dbbe
raw
history blame
8.47 kB
import spaces
import gc
import gradio as gr
import numpy as np
import os
from pathlib import Path
from diffusers import GGUFQuantizationConfig, HunyuanVideoPipeline, HunyuanVideoTransformer3DModel
from diffusers.utils import export_to_video
from huggingface_hub import snapshot_download
import torch
from PIL import Image
# Configuration
gc.collect()
torch.cuda.empty_cache()
torch.set_grad_enabled(False)
torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
# Load base model
model_id = "hunyuanvideo-community/HunyuanVideo"
base_path = f"/home/user/app/{model_id}"
os.makedirs(base_path, exist_ok=True)
snapshot_download(repo_id=model_id, local_dir=base_path)
# Load transformer
ckp_path = Path(base_path)
gguf_filename = "hunyuan-video-t2v-720p-Q4_0.gguf"
transformer_path = f"https://huggingface.co/city96/HunyuanVideo-gguf/blob/main/{gguf_filename}"
transformer = HunyuanVideoTransformer3DModel.from_single_file(
transformer_path,
quantization_config=GGUFQuantizationConfig(compute_dtype=torch.bfloat16),
torch_dtype=torch.bfloat16,
).to('cuda')
# Initialize pipeline
pipe = HunyuanVideoPipeline.from_pretrained(
ckp_path,
transformer=transformer,
torch_dtype=torch.float16
).to("cuda")
# Configure VAE
pipe.vae.enable_tiling()
pipe.vae.enable_slicing()
pipe.vae.eval()
# Available LORAs with display names
LORA_CHOICES = [
("stripe_v2.safetensors", "Stripe Style"),
("Top_Off.safetensors", "Top Off Effect"),
("huanyan_helper.safetensors", "Hunyuan Helper"),
("huanyan_helper_alpha.safetensors", "Hunyuan Alpha"),
("hunyuan-t-solo-v1.0.safetensors", "Solo Animation")
]
# Load all LORAs with hunyuanvideo-lora adapter
for weight_name, display_name in LORA_CHOICES:
pipe.load_lora_weights(
"Sergidev/TTV4ME",
weight_name=weight_name,
adapter_name=display_name.replace(" ", "_").lower(),
token=os.environ.get("HF_TOKEN")
)
# Memory cleanup
gc.collect()
torch.cuda.empty_cache()
MAX_SEED = np.iinfo(np.int32).max
MAX_IMAGE_SIZE = 1024
@spaces.GPU(duration=300)
def generate(
prompt,
image_input,
height,
width,
num_frames,
num_inference_steps,
seed_value,
fps,
selected_loras,
lora_weights,
progress=gr.Progress(track_tqdm=True)
):
# Validate image resolution
if image_input is not None:
img = Image.open(image_input)
if img.size != (width, height):
raise gr.Error(f"Image resolution {img.size} must match video resolution ({width}x{height})")
# Configure LORAs
active_adapters = [lora[1].replace(" ", "_").lower() for lora in LORA_CHOICES if lora[1] in selected_loras]
weights = [float(lora_weights[selected_loras.index(lora[1])]) for lora in LORA_CHOICES if lora[1] in selected_loras]
pipe.set_adapters(active_adapters, weights)
with torch.cuda.device(0):
if seed_value == -1:
seed_value = torch.randint(0, MAX_SEED, (1,)).item()
generator = torch.Generator('cuda').manual_seed(seed_value)
with torch.amp.autocast_mode.autocast('cuda', dtype=torch.bfloat16), torch.inference_mode(), torch.no_grad():
output = pipe(
prompt=prompt,
height=height,
width=width,
num_frames=num_frames,
num_inference_steps=num_inference_steps,
generator=generator,
).frames[0]
output_path = "output.mp4"
export_to_video(output, output_path, fps=fps)
torch.cuda.empty_cache()
gc.collect()
return output_path
def apply_preset(preset_name, *current_values):
if preset_name == "Higher Resolution":
return [608, 448, 24, 29, 12]
elif preset_name == "More Frames":
return [512, 320, 42, 27, 14]
return current_values
css = """
/* Existing CSS remains unchanged */
.lora-sliders {
margin-top: 15px;
border-top: 1px solid #444;
padding-top: 15px;
}
"""
with gr.Blocks(css=css, theme="dark") as demo:
with gr.Column(elem_id="col-container"):
gr.Markdown("# 🎬 Hunyuan Studio", elem_classes=["title"])
gr.Markdown(
"""Generate videos from text or images using multiple LoRA adapters.
Requires matching resolution between input image and output settings.""",
elem_classes=["description"]
)
with gr.Column(elem_classes=["prompt-container"]):
prompt = gr.Textbox(
label="Prompt",
placeholder="Enter text prompt or upload image below",
show_label=False,
elem_classes=["prompt-textbox"],
lines=3
)
image_input = gr.Image(type="filepath", label="Upload Image (Optional)")
with gr.Row():
run_button = gr.Button("🎨 Generate", variant="primary", size="lg")
with gr.Row(elem_classes=["preset-buttons"]):
preset_high_res = gr.Button("📺 Higher Resolution Preset")
preset_more_frames = gr.Button("🎞️ More Frames Preset")
with gr.Row():
result = gr.Video(label="Generated Video")
with gr.Accordion("⚙️ Advanced Settings", open=False):
seed = gr.Slider(
label="Seed (-1 for random)",
minimum=-1,
maximum=MAX_SEED,
step=1,
value=-1,
)
with gr.Row():
height = gr.Slider(
label="Height",
minimum=256,
maximum=MAX_IMAGE_SIZE,
step=16,
value=608,
)
width = gr.Slider(
label="Width",
minimum=256,
maximum=MAX_IMAGE_SIZE,
step=16,
value=448,
)
with gr.Row():
num_frames = gr.Slider(
label="Number of frames",
minimum=1.0,
maximum=257.0,
step=1,
value=24,
)
num_inference_steps = gr.Slider(
label="Inference steps",
minimum=1,
maximum=50,
step=1,
value=29,
)
fps = gr.Slider(
label="Frames per second",
minimum=1,
maximum=60,
step=1,
value=12,
)
with gr.Column(elem_classes=["lora-sliders"]):
gr.Markdown("### LoRA Adapters")
lora_checkboxes = gr.CheckboxGroup(
label="Select LoRAs",
choices=[display for (_, display) in LORA_CHOICES],
value=["Stripe Style", "Top Off Effect"]
)
lora_weight_sliders = []
for _, display_name in LORA_CHOICES:
lora_weight_sliders.append(
gr.Slider(
label=f"{display_name} Weight",
minimum=0.0,
maximum=1.0,
value=0.9 if "Stripe" in display_name else 0.8,
visible=False
)
)
# Event handling
run_button.click(
fn=generate,
inputs=[prompt, image_input, height, width, num_frames,
num_inference_steps, seed, fps, lora_checkboxes, lora_weight_sliders],
outputs=[result],
)
# Preset button handlers
preset_high_res.click(
fn=lambda: apply_preset("Higher Resolution"),
outputs=[height, width, num_frames, num_inference_steps, fps]
)
preset_more_frames.click(
fn=lambda: apply_preset("More Frames"),
outputs=[height, width, num_frames, num_inference_steps, fps]
)
# Show/hide LORA weight sliders based on checkbox selection
def toggle_lora_sliders(selected_loras):
updates = []
for lora in LORA_CHOICES:
updates.append(gr.update(visible=lora[1] in selected_loras))
return updates
lora_checkboxes.change(
fn=toggle_lora_sliders,
inputs=lora_checkboxes,
outputs=lora_weight_sliders
)