jbilcke-hf's picture
jbilcke-hf HF Staff
Fix for #22
883eb72
"""
Preview service for Video Model Studio
Handles the video generation logic and model integration
"""
import logging
import tempfile
import traceback
import random
from pathlib import Path
from typing import Dict, Any, List, Optional, Tuple, Callable
import time
from vms.config import (
OUTPUT_PATH, STORAGE_PATH, MODEL_TYPES, TRAINING_PATH,
DEFAULT_PROMPT_PREFIX, MODEL_VERSIONS
)
from vms.utils import format_time
logger = logging.getLogger(__name__)
logger.setLevel(logging.INFO)
class PreviewingService:
"""Handles the video generation logic and model integration"""
def __init__(self):
"""Initialize the preview service"""
pass
def find_latest_lora_weights(self) -> Optional[str]:
"""Find the latest LoRA weights file"""
try:
lora_path = OUTPUT_PATH / "pytorch_lora_weights.safetensors"
if lora_path.exists():
return str(lora_path)
# If not found in the expected location, try to find in checkpoints
checkpoints = list(OUTPUT_PATH.glob("finetrainers_step_*"))
has_checkpoints = len(checkpoints) > 0
if not checkpoints:
return None
latest_checkpoint = max(checkpoints, key=lambda x: int(x.name.split("-")[1]))
lora_path = latest_checkpoint / "pytorch_lora_weights.safetensors"
if lora_path.exists():
return str(lora_path)
return None
except Exception as e:
logger.error(f"Error finding LoRA weights: {e}")
return None
def get_model_versions(self, model_type: str) -> Dict[str, Dict[str, str]]:
"""Get available model versions for the given model type"""
return MODEL_VERSIONS.get(model_type, {})
def generate_video(
self,
model_type: str,
model_version: str,
prompt: str,
negative_prompt: str,
prompt_prefix: str,
width: int,
height: int,
num_frames: int,
guidance_scale: float,
flow_shift: float,
lora_scale: float,
inference_steps: int,
seed: int = -1,
enable_cpu_offload: bool = True,
fps: int = 16,
conditioning_image: Optional[str] = None
) -> Tuple[Optional[str], str, str]:
"""Generate a video using the trained model"""
try:
log_messages = []
print("generate_video")
def log(msg: str):
log_messages.append(msg)
logger.info(msg)
# Return updated log string for UI updates
return "\n".join(log_messages)
# Find latest LoRA weights if lora_scale > 0
lora_path = None
using_lora = lora_scale > 0
if using_lora:
lora_path = self.find_latest_lora_weights()
if not lora_path:
return None, "Error: No LoRA weights found", log("Error: No LoRA weights found in output directory")
log(f"Using LoRA weights with scale {lora_scale}")
else:
log("Using original model without LoRA weights")
# Add prefix to prompt
if prompt_prefix and not prompt.startswith(prompt_prefix):
full_prompt = f"{prompt_prefix}{prompt}"
else:
full_prompt = prompt
# Create correct num_frames (should be 8*k + 1)
adjusted_num_frames = ((num_frames - 1) // 8) * 8 + 1
if adjusted_num_frames != num_frames:
log(f"Adjusted number of frames from {num_frames} to {adjusted_num_frames} to match model requirements")
num_frames = adjusted_num_frames
# Get model type (internal name)
internal_model_type = MODEL_TYPES.get(model_type)
if not internal_model_type:
return None, f"Error: Invalid model type {model_type}", log(f"Error: Invalid model type {model_type}")
# Check if model version is valid
# This section uses model_version directly from parameter
if model_version:
# Verify that the specified model_version exists in our versions
versions = self.get_model_versions(internal_model_type)
if model_version not in versions:
log(f"Warning: Specified model version '{model_version}' is not recognized")
# Fall back to default version for this model
if len(versions) > 0:
model_version = next(iter(versions.keys()))
log(f"Using default model version instead: {model_version}")
else:
log(f"Using specified model version: {model_version}")
else:
# No model version specified, use default
versions = self.get_model_versions(internal_model_type)
if len(versions) > 0:
model_version = next(iter(versions.keys()))
log(f"No model version specified, using default: {model_version}")
else:
# Fall back to hardcoded defaults if no versions defined
if internal_model_type == "wan":
model_version = "Wan-AI/Wan2.1-T2V-1.3B-Diffusers"
elif internal_model_type == "ltx_video":
model_version = "Lightricks/LTX-Video"
elif internal_model_type == "hunyuan_video":
model_version = "hunyuanvideo-community/HunyuanVideo"
log(f"No versions defined for model type, using default: {model_version}")
# Check if this is an image-to-video model but no image was provided
model_version_info = versions.get(model_version, {})
if model_version_info.get("type") == "image-to-video" and not conditioning_image:
return None, "Error: This model requires a conditioning image", log("Error: This model version requires a conditioning image but none was provided")
log(f"Generating video with model type: {internal_model_type}")
log(f"Using model version: {model_version}")
if using_lora and lora_path:
log(f"Using LoRA weights from: {lora_path}")
log(f"Resolution: {width}x{height}, Frames: {num_frames}, FPS: {fps}")
log(f"Guidance Scale: {guidance_scale}, Flow Shift: {flow_shift}, LoRA Scale: {lora_scale if using_lora else 0}")
log(f"Generation Seed: {seed}")
#log(f"Prompt: {full_prompt}")
#log(f"Negative Prompt: {negative_prompt}")
# Import required components based on model type
if internal_model_type == "wan":
return self.generate_wan_video(
full_prompt, negative_prompt, width, height, num_frames,
guidance_scale, flow_shift, lora_path, lora_scale,
inference_steps, seed, enable_cpu_offload, fps, log,
model_version, conditioning_image
)
elif internal_model_type == "ltx_video":
return self.generate_ltx_video(
full_prompt, negative_prompt, width, height, num_frames,
guidance_scale, flow_shift, lora_path, lora_scale,
inference_steps, seed, enable_cpu_offload, fps, log,
model_version, conditioning_image
)
elif internal_model_type == "hunyuan_video":
return self.generate_hunyuan_video(
full_prompt, negative_prompt, width, height, num_frames,
guidance_scale, flow_shift, lora_path, lora_scale,
inference_steps, seed, enable_cpu_offload, fps, log,
model_version, conditioning_image
)
else:
return None, f"Error: Unsupported model type {internal_model_type}", log(f"Error: Unsupported model type {internal_model_type}")
except Exception as e:
logger.exception("Error generating video")
return None, f"Error: {str(e)}", f"Exception occurred: {str(e)}"
def generate_wan_video(
self,
prompt: str,
negative_prompt: str,
width: int,
height: int,
num_frames: int,
guidance_scale: float,
flow_shift: float,
lora_path: str,
lora_scale: float,
inference_steps: int,
seed: int = -1,
enable_cpu_offload: bool = True,
fps: int = 16,
log_fn: Callable = print,
model_version: str = "Wan-AI/Wan2.1-T2V-1.3B-Diffusers",
conditioning_image: Optional[str] = None
) -> Tuple[Optional[str], str, str]:
"""Generate video using Wan model"""
try:
import torch
import numpy as np
from diffusers import AutoencoderKLWan, WanPipeline
from diffusers.schedulers.scheduling_unipc_multistep import UniPCMultistepScheduler
from diffusers.utils import export_to_video
from PIL import Image
import os
start_time = torch.cuda.Event(enable_timing=True)
end_time = torch.cuda.Event(enable_timing=True)
print("Initializing wan generation..")
log_fn("Importing Wan model components...")
# Set up random seed
if seed == -1:
seed = random.randint(0, 2**32 - 1)
log_fn(f"Using randomly generated seed: {seed}")
# Set random seeds for reproducibility
random.seed(seed)
np.random.seed(seed)
torch.manual_seed(seed)
generator = torch.Generator(device="cuda")
generator = generator.manual_seed(seed)
log_fn(f"Loading VAE from {model_version}...")
vae = AutoencoderKLWan.from_pretrained(model_version, subfolder="vae", torch_dtype=torch.float32)
log_fn(f"Loading transformer from {model_version}...")
pipe = WanPipeline.from_pretrained(model_version, vae=vae, torch_dtype=torch.bfloat16)
log_fn(f"Configuring scheduler with flow_shift={flow_shift}...")
pipe.scheduler = UniPCMultistepScheduler.from_config(
pipe.scheduler.config,
flow_shift=flow_shift
)
log_fn("Moving pipeline to CUDA device...")
pipe.to("cuda")
if enable_cpu_offload:
log_fn("Enabling model CPU offload...")
pipe.enable_model_cpu_offload()
# Apply LoRA weights if using them
if lora_scale > 0 and lora_path:
log_fn(f"Loading LoRA weights from {lora_path} with lora scale {lora_scale}...")
pipe.load_lora_weights(lora_path)
else:
log_fn("Using base model without LoRA weights")
# Create temporary file for the output
with tempfile.NamedTemporaryFile(suffix='.mp4', delete=False) as temp_file:
output_path = temp_file.name
log_fn("Starting video generation...")
start_time.record()
# Check if this is an image-to-video model
is_i2v = "I2V" in model_version
if is_i2v and conditioning_image:
log_fn(f"Loading conditioning image from {conditioning_image}...")
image = Image.open(conditioning_image).convert("RGB")
image = image.resize((width, height))
log_fn("Generating video with image conditioning...")
output = pipe(
prompt=prompt,
negative_prompt=negative_prompt,
image=image,
height=height,
width=width,
num_frames=num_frames,
guidance_scale=guidance_scale,
num_inference_steps=inference_steps,
cross_attention_kwargs={"scale": lora_scale},
generator=generator,
).frames[0]
else:
log_fn("Generating video with text-only conditioning...")
output = pipe(
prompt=prompt,
negative_prompt=negative_prompt,
height=height,
width=width,
num_frames=num_frames,
guidance_scale=guidance_scale,
num_inference_steps=inference_steps,
generator=generator,
).frames[0]
end_time.record()
torch.cuda.synchronize()
generation_time = start_time.elapsed_time(end_time) / 1000 # Convert to seconds
log_fn(f"Video generation completed in {format_time(generation_time)}")
log_fn(f"Exporting video to {output_path}...")
export_to_video(output, output_path, fps=fps)
log_fn("Video generation and export completed successfully!")
# Clean up CUDA memory
pipe = None
torch.cuda.empty_cache()
return output_path, "Video generated successfully!", log_fn(f"Generation completed in {format_time(generation_time)}")
except Exception as e:
traceback.print_exc()
log_fn(f"Error generating video with Wan: {str(e)}")
# Clean up CUDA memory
torch.cuda.empty_cache()
return None, f"Error: {str(e)}", log_fn(f"Exception occurred: {str(e)}")
def generate_ltx_video(
self,
prompt: str,
negative_prompt: str,
width: int,
height: int,
num_frames: int,
guidance_scale: float,
flow_shift: float,
lora_path: str,
lora_scale: float,
inference_steps: int,
seed: int = -1,
enable_cpu_offload: bool = True,
fps: int = 16,
log_fn: Callable = print,
model_version: str = "Lightricks/LTX-Video",
conditioning_image: Optional[str] = None
) -> Tuple[Optional[str], str, str]:
"""Generate video using LTX model"""
try:
import torch
import numpy as np
from diffusers import LTXPipeline
from diffusers.utils import export_to_video
from PIL import Image
start_time = torch.cuda.Event(enable_timing=True)
end_time = torch.cuda.Event(enable_timing=True)
# Set up random seed
if seed == -1:
seed = random.randint(0, 2**32 - 1)
log_fn(f"Using randomly generated seed: {seed}")
# Set random seeds for reproducibility
random.seed(seed)
np.random.seed(seed)
torch.manual_seed(seed)
generator = torch.Generator(device="cuda")
generator = generator.manual_seed(seed)
log_fn("Importing LTX model components...")
log_fn(f"Loading pipeline from {model_version}...")
pipe = LTXPipeline.from_pretrained(model_version, torch_dtype=torch.bfloat16)
log_fn("Moving pipeline to CUDA device...")
pipe.to("cuda")
if enable_cpu_offload:
log_fn("Enabling model CPU offload...")
pipe.enable_model_cpu_offload()
# Apply LoRA weights if using them
if lora_scale > 0 and lora_path:
log_fn(f"Loading LoRA weights from {lora_path} with lora scale {lora_scale}...")
pipe.load_lora_weights(lora_path)
else:
log_fn("Using base model without LoRA weights")
# Create temporary file for the output
with tempfile.NamedTemporaryFile(suffix='.mp4', delete=False) as temp_file:
output_path = temp_file.name
log_fn("Starting video generation...")
start_time.record()
# LTX doesn't currently support image conditioning in the standard way
video = pipe(
prompt=prompt,
negative_prompt=negative_prompt,
height=height,
width=width,
num_frames=num_frames,
guidance_scale=guidance_scale,
decode_timestep=0.03,
decode_noise_scale=0.025,
num_inference_steps=inference_steps,
cross_attention_kwargs={"scale": lora_scale},
generator=generator,
).frames[0]
end_time.record()
torch.cuda.synchronize()
generation_time = start_time.elapsed_time(end_time) / 1000 # Convert to seconds
log_fn(f"Video generation completed in {format_time(generation_time)}")
log_fn(f"Exporting video to {output_path}...")
export_to_video(video, output_path, fps=fps)
log_fn("Video generation and export completed successfully!")
# Clean up CUDA memory
pipe = None
torch.cuda.empty_cache()
return output_path, "Video generated successfully!", log_fn(f"Generation completed in {format_time(generation_time)}")
except Exception as e:
log_fn(f"Error generating video with LTX: {str(e)}")
# Clean up CUDA memory
torch.cuda.empty_cache()
return None, f"Error: {str(e)}", log_fn(f"Exception occurred: {str(e)}")
def generate_hunyuan_video(
self,
prompt: str,
negative_prompt: str,
width: int,
height: int,
num_frames: int,
guidance_scale: float,
flow_shift: float,
lora_path: str,
lora_scale: float,
inference_steps: int,
seed: int = -1,
enable_cpu_offload: bool = True,
fps: int = 16,
log_fn: Callable = print,
model_version: str = "hunyuanvideo-community/HunyuanVideo",
conditioning_image: Optional[str] = None
) -> Tuple[Optional[str], str, str]:
"""Generate video using HunyuanVideo model"""
try:
import torch
import numpy as np
from diffusers import HunyuanVideoPipeline, HunyuanVideoTransformer3DModel, AutoencoderKLHunyuanVideo
from diffusers.utils import export_to_video
start_time = torch.cuda.Event(enable_timing=True)
end_time = torch.cuda.Event(enable_timing=True)
# Set up random seed
if seed == -1:
seed = random.randint(0, 2**32 - 1)
log_fn(f"Using randomly generated seed: {seed}")
# Set random seeds for reproducibility
random.seed(seed)
np.random.seed(seed)
torch.manual_seed(seed)
generator = torch.Generator(device="cuda")
generator = generator.manual_seed(seed)
log_fn("Importing HunyuanVideo model components...")
log_fn(f"Loading transformer from {model_version}...")
transformer = HunyuanVideoTransformer3DModel.from_pretrained(
model_version,
subfolder="transformer",
torch_dtype=torch.bfloat16
)
log_fn(f"Loading pipeline from {model_version}...")
pipe = HunyuanVideoPipeline.from_pretrained(
model_version,
transformer=transformer,
torch_dtype=torch.float16
)
log_fn("Enabling VAE tiling for better memory usage...")
pipe.vae.enable_tiling()
log_fn("Moving pipeline to CUDA device...")
pipe.to("cuda")
if enable_cpu_offload:
log_fn("Enabling model CPU offload...")
pipe.enable_model_cpu_offload()
# Apply LoRA weights if using them
if lora_scale > 0 and lora_path:
log_fn(f"Loading LoRA weights from {lora_path} with lora scale {lora_scale}...")
pipe.load_lora_weights(lora_path)
else:
log_fn("Using base model without LoRA weights")
# Create temporary file for the output
with tempfile.NamedTemporaryFile(suffix='.mp4', delete=False) as temp_file:
output_path = temp_file.name
log_fn("Starting video generation...")
start_time.record()
# Make sure negative_prompt is a list or None
neg_prompt = [negative_prompt] if negative_prompt else None
output = pipe(
prompt=prompt,
negative_prompt=neg_prompt,
height=height,
width=width,
num_frames=num_frames,
guidance_scale=guidance_scale,
true_cfg_scale=1.0,
num_inference_steps=inference_steps,
cross_attention_kwargs={"scale": lora_scale},
generator=generator,
).frames[0]
end_time.record()
torch.cuda.synchronize()
generation_time = start_time.elapsed_time(end_time) / 1000 # Convert to seconds
log_fn(f"Video generation completed in {format_time(generation_time)}")
log_fn(f"Exporting video to {output_path}...")
export_to_video(output, output_path, fps=fps)
log_fn("Video generation and export completed successfully!")
# Clean up CUDA memory
pipe = None
torch.cuda.empty_cache()
return output_path, "Video generated successfully!", log_fn(f"Generation completed in {format_time(generation_time)}")
except Exception as e:
log_fn(f"Error generating video with HunyuanVideo: {str(e)}")
# Clean up CUDA memory
torch.cuda.empty_cache()
return None, f"Error: {str(e)}", log_fn(f"Exception occurred: {str(e)}")