Framepack-H111 / wan /image2video.py
rahul7star's picture
Upload 303 files
e0336bc verified
# Modified from official implementation
# Original source:
# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
import logging
import os
import random
import sys
from typing import Optional, Union
import cv2
import numpy as np
import torch
import torchvision.transforms.functional as TF
from tqdm import tqdm
from accelerate import Accelerator, init_empty_weights
from modules.scheduling_flow_match_discrete import FlowMatchDiscreteScheduler
from utils.safetensors_utils import load_safetensors
# from .distributed.fsdp import shard_model
from .modules.clip import CLIPModel
from .modules.model import WanModel, load_wan_model
from .modules.t5 import T5EncoderModel
from .modules.vae import WanVAE
from .utils.fm_solvers import FlowDPMSolverMultistepScheduler, get_sampling_sigmas, retrieve_timesteps
from .utils.fm_solvers_unipc import FlowUniPCMultistepScheduler
from utils.device_utils import clean_memory_on_device, synchronize_device
import logging
logger = logging.getLogger(__name__)
logging.basicConfig(level=logging.INFO)
class WanI2V:
def __init__(
self,
config,
checkpoint_dir,
device_id=0,
rank=0,
t5_fsdp=False,
dit_fsdp=False,
use_usp=False,
t5_cpu=False,
init_on_cpu=True,
device=None,
dit_dtype=None,
dit_weight_dtype=None,
dit_path=None,
dit_attn_mode=None,
t5_path=None,
clip_path=None,
t5_fp8=False,
):
r"""
Initializes the image-to-video generation model components.
Args:
config (EasyDict):
Object containing model parameters initialized from config.py
checkpoint_dir (`str`):
Path to directory containing model checkpoints
device_id (`int`, *optional*, defaults to 0) **IGNORED**:
Id of target GPU device
rank (`int`, *optional*, defaults to 0) **IGNORED**:
Process rank for distributed training
t5_fsdp (`bool`, *optional*, defaults to False) **IGNORED**:
Enable FSDP sharding for T5 model
dit_fsdp (`bool`, *optional*, defaults to False) **IGNORED**:
Enable FSDP sharding for DiT model
use_usp (`bool`, *optional*, defaults to False) **IGNORED**:
Enable distribution strategy of USP.
t5_cpu (`bool`, *optional*, defaults to False) **IGNORED**:
Whether to place T5 model on CPU. Only works without t5_fsdp.
init_on_cpu (`bool`, *optional*, defaults to True) **IGNORED**:
Enable initializing Transformer Model on CPU. Only works without FSDP or USP.
device (`torch.device`, *optional*, defaults to None):
Device to place the model on. If None, use the default device (cuda)
dtype (`torch.dtype`, *optional*, defaults to None):
Data type for DiT model parameters. If None, use the default parameter data type from config
dit_path (`str`, *optional*, defaults to None):
Path to DiT model checkpoint. checkpoint_dir is used if None.
dit_attn_mode (`str`, *optional*, defaults to None):
Attention mode for DiT model. If None, use "torch" attention mode.
t5_path (`str`, *optional*, defaults to None):
Path to T5 model checkpoint. checkpoint_dir is used if None.
clip_path (`str`, *optional*, defaults to None):
Path to CLIP model checkpoint. checkpoint_dir is used if None.
t5_fp8 (`bool`, *optional*, defaults to False):
Enable FP8 quantization for T5 model
"""
self.device = device if device is not None else torch.device("cuda")
self.config = config
self.rank = rank
self.t5_cpu = t5_cpu
self.t5_fp8 = t5_fp8
self.num_train_timesteps = config.num_train_timesteps
self.param_dtype = config.param_dtype
# shard_fn = partial(shard_model, device_id=device_id)
checkpoint_path = None if checkpoint_dir is None else os.path.join(checkpoint_dir, config.t5_checkpoint)
tokenizer_path = None if checkpoint_dir is None else os.path.join(checkpoint_dir, config.t5_tokenizer)
self.text_encoder = T5EncoderModel(
text_len=config.text_len,
dtype=config.t5_dtype,
device=device,
checkpoint_path=checkpoint_path,
tokenizer_path=tokenizer_path,
weight_path=t5_path,
fp8=t5_fp8,
# shard_fn=shard_fn if t5_fsdp else None,
)
self.vae_stride = config.vae_stride
self.patch_size = config.patch_size
self.checkpoint_dir = checkpoint_dir
self.dit_path = dit_path
self.dit_dtype = dit_dtype # if dit_dtype is not None else config.param_dtype
self.dit_weight_dtype = dit_weight_dtype
self.dit_attn_mode = dit_attn_mode
self.clip_path = clip_path
self.sample_neg_prompt = config.sample_neg_prompt
def generate(
self,
accelerator: Accelerator,
merge_lora: Optional[callable],
fp8_scaled: bool,
input_prompt,
img,
size=(1280, 720),
frame_num=81,
shift=5.0,
sample_solver="unipc",
sampling_steps=40,
guide_scale=5.0,
n_prompt="",
seed=-1,
blocks_to_swap=0,
vae: WanVAE = None,
):
r"""
Generates video frames from input image and text prompt using diffusion process.
Args:
input_prompt (`str`):
Text prompt for content generation.
img (PIL.Image.Image):
Input image tensor. Shape: [3, H, W]
max_area (`int`, *optional*, defaults to 720*1280):
Maximum pixel area for latent space calculation. Controls video resolution scaling
frame_num (`int`, *optional*, defaults to 81):
How many frames to sample from a video. The number should be 4n+1
shift (`float`, *optional*, defaults to 5.0):
Noise schedule shift parameter. Affects temporal dynamics
[NOTE]: If you want to generate a 480p video, it is recommended to set the shift value to 3.0.
sample_solver (`str`, *optional*, defaults to 'unipc'):
Solver used to sample the video.
sampling_steps (`int`, *optional*, defaults to 40):
Number of diffusion sampling steps. Higher values improve quality but slow generation
guide_scale (`float`, *optional*, defaults 5.0):
Classifier-free guidance scale. Controls prompt adherence vs. creativity
n_prompt (`str`, *optional*, defaults to ""):
Negative prompt for content exclusion. If not given, use `config.sample_neg_prompt`
seed (`int`, *optional*, defaults to -1):
Random seed for noise generation. If -1, use random seed.
blocks_to_swap (`int`, *optional*, defaults to 0):
Number of blocks to swap (offload) to CPU. If 0, no blocks are offloaded.
Returns:
torch.Tensor:
Generated video frames tensor. Dimensions: (C, N H, W) where:
- C: Color channels (3 for RGB)
- N: Number of frames (81)
- H: Frame height (from size)
- W: Frame width from size)
"""
max_area = size[0] * size[1]
# save original image as numpy array
img_cv2 = np.array(img) # PIL to numpy
img_cv2 = cv2.cvtColor(img_cv2, cv2.COLOR_BGR2RGB)
img = TF.to_tensor(img).sub_(0.5).div_(0.5).to(self.device) # -1 to 1
F = frame_num # number of frames
h, w = img.shape[1:]
aspect_ratio = h / w
lat_h = round(np.sqrt(max_area * aspect_ratio) // self.vae_stride[1] // self.patch_size[1] * self.patch_size[1])
lat_w = round(np.sqrt(max_area / aspect_ratio) // self.vae_stride[2] // self.patch_size[2] * self.patch_size[2])
h = lat_h * self.vae_stride[1]
w = lat_w * self.vae_stride[2]
lat_f = (F - 1) // self.vae_stride[0] + 1 # size of latent frames
max_seq_len = lat_f * lat_h * lat_w // (self.patch_size[1] * self.patch_size[2])
# set seed
seed = seed if seed >= 0 else random.randint(0, sys.maxsize)
seed_g = torch.Generator(device=self.device)
seed_g.manual_seed(seed)
# Generate noise for the required number of frames only
noise = torch.randn(16, lat_f, lat_h, lat_w, dtype=torch.float32, generator=seed_g, device=self.device)
if n_prompt == "":
n_prompt = self.sample_neg_prompt
# preprocess
self.text_encoder.model.to(self.device)
with torch.no_grad():
if self.t5_fp8:
with accelerator.autocast():
context = self.text_encoder([input_prompt], self.device)
context_null = self.text_encoder([n_prompt], self.device)
else:
context = self.text_encoder([input_prompt], self.device)
context_null = self.text_encoder([n_prompt], self.device)
del self.text_encoder
clean_memory_on_device(self.device)
# load CLIP model
checkpoint_path = None if self.checkpoint_dir is None else os.path.join(self.checkpoint_dir, self.config.clip_checkpoint)
tokenizer_path = None if self.checkpoint_dir is None else os.path.join(self.checkpoint_dir, self.config.clip_tokenizer)
clip = CLIPModel(
dtype=self.config.clip_dtype,
device=self.device,
checkpoint_path=checkpoint_path,
tokenizer_path=tokenizer_path,
weight_path=self.clip_path,
)
clip.model.to(self.device)
logger.info(f"Encoding image to CLIP context")
# use torch.amp.autocast istead of accelerator.autocast, becuase CLIP dtype is not bfloat16
with torch.amp.autocast(device_type=self.device.type, dtype=torch.float16), torch.no_grad():
clip_context = clip.visual([img[:, None, :, :]])
logger.info(f"Encoding complete")
del clip
clean_memory_on_device(self.device)
# y should be encoded with 81 frames, and trim to lat_f frames? encoding F frames causes invalid results?
logger.info(f"Encoding image to latent space")
vae.to_device(self.device)
# resize image for the first frame. INTER_AREA is the best for downsampling
interpolation = cv2.INTER_AREA if h < img_cv2.shape[0] else cv2.INTER_CUBIC
img_resized = cv2.resize(img_cv2, (w, h), interpolation=interpolation)
img_resized = cv2.cvtColor(img_resized, cv2.COLOR_BGR2RGB)
img_resized = TF.to_tensor(img_resized).sub_(0.5).div_(0.5).to(self.device) # -1 to 1, CHW
img_resized = img_resized.unsqueeze(1) # CFHW
# Create mask for the required number of frames
msk = torch.ones(1, F, lat_h, lat_w, device=self.device)
msk[:, 1:] = 0
msk = torch.concat([torch.repeat_interleave(msk[:, 0:1], repeats=4, dim=1), msk[:, 1:]], dim=1)
msk = msk.view(1, msk.shape[1] // 4, 4, lat_h, lat_w)
msk = msk.transpose(1, 2)[0]
with accelerator.autocast(), torch.no_grad():
# Zero padding for the required number of frames only
padding_frames = F - 1 # The first frame is the input image
img_resized = torch.concat([img_resized, torch.zeros(3, padding_frames, h, w, device=self.device)], dim=1)
y = vae.encode([img_resized])[0]
y = y[:, :lat_f] # may be not needed
y = torch.concat([msk, y])
logger.info(f"Encoding complete")
vae.to_device("cpu")
clean_memory_on_device(self.device)
# load DiT model
loading_device = "cpu"
if blocks_to_swap == 0 and merge_lora is None and not fp8_scaled:
loading_device = self.device
loading_weight_dtype = self.dit_weight_dtype
if fp8_scaled or merge_lora is not None:
loading_weight_dtype = self.dit_dtype # load as-is
# set fp8_scaled to False, because we optimize the model after merging LoRA
# TODO state dict based LoRA merge
self.model: WanModel = load_wan_model(
self.config,
True,
self.device,
self.dit_path,
self.dit_attn_mode,
False,
loading_device,
loading_weight_dtype,
False,
)
if merge_lora is not None:
# merge LoRA to the model, cast and move to the device
merge_lora(self.model)
if fp8_scaled:
state_dict = self.model.state_dict()
move_to_device = blocks_to_swap == 0 # if blocks_to_swap > 0, we will keep the model on CPU
state_dict = self.model.fp8_optimization(state_dict, self.device, move_to_device)
info = self.model.load_state_dict(state_dict, strict=True, assign=True)
logger.info(f"Loaded FP8 optimized weights: {info}")
if blocks_to_swap == 0:
self.model.to(self.device) # make sure all parameters are on the right device
else:
target_dtype = None
target_device = None
if self.dit_weight_dtype is not None: # in case of args.fp8 (not fp8_scaled)
logger.info(f"Convert model to {self.dit_weight_dtype}")
target_dtype = self.dit_weight_dtype
if blocks_to_swap == 0:
logger.info(f"Move model to device: {self.device}")
target_device = self.device
self.model.to(target_device, target_dtype)
if blocks_to_swap > 0:
logger.info(f"Enable swap {blocks_to_swap} blocks to CPU from device: {self.device}")
self.model.enable_block_swap(blocks_to_swap, self.device, supports_backward=False)
self.model.move_to_device_except_swap_blocks(self.device)
self.model.prepare_block_swap_before_forward()
else:
# make sure the model is on the right device
self.model.to(self.device)
self.model.eval().requires_grad_(False)
clean_memory_on_device(self.device)
# evaluation mode
with torch.no_grad():
if sample_solver == "unipc":
sample_scheduler = FlowUniPCMultistepScheduler(
num_train_timesteps=self.num_train_timesteps, shift=1, use_dynamic_shifting=False
)
sample_scheduler.set_timesteps(sampling_steps, device=self.device, shift=shift)
timesteps = sample_scheduler.timesteps
elif sample_solver == "dpm++":
sample_scheduler = FlowDPMSolverMultistepScheduler(
num_train_timesteps=self.num_train_timesteps, shift=1, use_dynamic_shifting=False
)
sampling_sigmas = get_sampling_sigmas(sampling_steps, shift)
timesteps, _ = retrieve_timesteps(sample_scheduler, device=self.device, sigmas=sampling_sigmas)
elif sample_solver == "vanilla":
sample_scheduler = FlowMatchDiscreteScheduler(num_train_timesteps=self.num_train_timesteps, shift=shift)
sample_scheduler.set_timesteps(sampling_steps, device=self.device)
timesteps = sample_scheduler.timesteps
org_step = sample_scheduler.step
def step_wrapper(
model_output: torch.Tensor,
timestep: Union[int, torch.Tensor],
sample: torch.Tensor,
return_dict: bool = True,
generator=None,
):
return org_step(model_output, timestep, sample, return_dict=return_dict)
sample_scheduler.step = step_wrapper
else:
raise NotImplementedError("Unsupported solver.")
# sample videos
latent = noise # on device
del noise
arg_c = {
"context": [context[0]],
"clip_fea": clip_context,
"seq_len": max_seq_len,
"y": [y],
}
arg_null = {
"context": context_null,
"clip_fea": clip_context,
"seq_len": max_seq_len,
"y": [y],
}
# self.model.to(self.device)
for _, t in enumerate(tqdm(timesteps)):
latent_model_input = [latent.to(self.device)]
latent = latent.to("cpu")
timestep = [t]
timestep = torch.stack(timestep).to(self.device)
with accelerator.autocast():
noise_pred_cond = self.model(latent_model_input, t=timestep, **arg_c)[0].to("cpu")
noise_pred_uncond = self.model(latent_model_input, t=timestep, **arg_null)[0].to("cpu")
latent_model_input = None
noise_pred = noise_pred_uncond + guide_scale * (noise_pred_cond - noise_pred_uncond)
temp_x0 = sample_scheduler.step(
noise_pred.unsqueeze(0), t, latent.unsqueeze(0), return_dict=False, generator=seed_g
)[0]
latent = temp_x0.squeeze(0)
# x0 = [latent.to(self.device)]
del latent_model_input, timestep
del sample_scheduler
del self.model
synchronize_device(self.device)
clean_memory_on_device(self.device)
return latent