support motion controller

diffsynth/models/wan_video_controlnet.py

@@ -0,0 +1,204 @@
+import torch
+import torch.nn as nn
+from typing import Tuple, Optional
+from einops import rearrange
+from .wan_video_dit import DiTBlock, precompute_freqs_cis_3d, MLP, sinusoidal_embedding_1d
+from .utils import hash_state_dict_keys
+
+
+
+class WanControlNetModel(torch.nn.Module):
+    def __init__(
+        self,
+        dim: int,
+        in_dim: int,
+        ffn_dim: int,
+        out_dim: int,
+        text_dim: int,
+        freq_dim: int,
+        eps: float,
+        patch_size: Tuple[int, int, int],
+        num_heads: int,
+        num_layers: int,
+        has_image_input: bool,
+    ):
+        super().__init__()
+        self.dim = dim
+        self.freq_dim = freq_dim
+        self.has_image_input = has_image_input
+        self.patch_size = patch_size
+
+        self.patch_embedding = nn.Conv3d(
+            in_dim, dim, kernel_size=patch_size, stride=patch_size)
+        self.text_embedding = nn.Sequential(
+            nn.Linear(text_dim, dim),
+            nn.GELU(approximate='tanh'),
+            nn.Linear(dim, dim)
+        )
+        self.time_embedding = nn.Sequential(
+            nn.Linear(freq_dim, dim),
+            nn.SiLU(),
+            nn.Linear(dim, dim)
+        )
+        self.time_projection = nn.Sequential(
+            nn.SiLU(), nn.Linear(dim, dim * 6))
+        self.blocks = nn.ModuleList([
+            DiTBlock(has_image_input, dim, num_heads, ffn_dim, eps)
+            for _ in range(num_layers)
+        ])
+        head_dim = dim // num_heads
+        self.freqs = precompute_freqs_cis_3d(head_dim)
+
+        if has_image_input:
+            self.img_emb = MLP(1280, dim)  # clip_feature_dim = 1280
+            
+        self.controlnet_conv_in = torch.nn.Conv3d(in_channels=in_dim, out_channels=in_dim, kernel_size=1)
+        self.controlnet_blocks = torch.nn.ModuleList([
+            torch.nn.Linear(dim, dim, bias=False)
+            for _ in range(num_layers)
+        ])
+
+    def patchify(self, x: torch.Tensor):
+        x = self.patch_embedding(x)
+        grid_size = x.shape[2:]
+        x = rearrange(x, 'b c f h w -> b (f h w) c').contiguous()
+        return x, grid_size  # x, grid_size: (f, h, w)
+
+    def unpatchify(self, x: torch.Tensor, grid_size: torch.Tensor):
+        return rearrange(
+            x, 'b (f h w) (x y z c) -> b c (f x) (h y) (w z)',
+            f=grid_size[0], h=grid_size[1], w=grid_size[2], 
+            x=self.patch_size[0], y=self.patch_size[1], z=self.patch_size[2]
+        )
+
+    def forward(self,
+                x: torch.Tensor,
+                timestep: torch.Tensor,
+                context: torch.Tensor,
+                clip_feature: Optional[torch.Tensor] = None,
+                y: Optional[torch.Tensor] = None,
+                controlnet_conditioning: Optional[torch.Tensor] = None,
+                use_gradient_checkpointing: bool = False,
+                use_gradient_checkpointing_offload: bool = False,
+                **kwargs,
+                ):
+        t = self.time_embedding(
+            sinusoidal_embedding_1d(self.freq_dim, timestep))
+        t_mod = self.time_projection(t).unflatten(1, (6, self.dim))
+        context = self.text_embedding(context)
+        
+        if self.has_image_input:
+            x = torch.cat([x, y], dim=1)  # (b, c_x + c_y, f, h, w)
+            clip_embdding = self.img_emb(clip_feature)
+            context = torch.cat([clip_embdding, context], dim=1)
+            
+        x = x + self.controlnet_conv_in(controlnet_conditioning)
+        x, (f, h, w) = self.patchify(x)
+        
+        freqs = torch.cat([
+            self.freqs[0][:f].view(f, 1, 1, -1).expand(f, h, w, -1),
+            self.freqs[1][:h].view(1, h, 1, -1).expand(f, h, w, -1),
+            self.freqs[2][:w].view(1, 1, w, -1).expand(f, h, w, -1)
+        ], dim=-1).reshape(f * h * w, 1, -1).to(x.device)
+        
+        def create_custom_forward(module):
+            def custom_forward(*inputs):
+                return module(*inputs)
+            return custom_forward
+
+        res_stack = []
+        for block in self.blocks:
+            if self.training and use_gradient_checkpointing:
+                if use_gradient_checkpointing_offload:
+                    with torch.autograd.graph.save_on_cpu():
+                        x = torch.utils.checkpoint.checkpoint(
+                            create_custom_forward(block),
+                            x, context, t_mod, freqs,
+                            use_reentrant=False,
+                        )
+                else:
+                    x = torch.utils.checkpoint.checkpoint(
+                        create_custom_forward(block),
+                        x, context, t_mod, freqs,
+                        use_reentrant=False,
+                    )
+            else:
+                x = block(x, context, t_mod, freqs)
+            res_stack.append(x)
+        
+        controlnet_res_stack = [block(res) for block, res in zip(self.controlnet_blocks, res_stack)]
+        return controlnet_res_stack
+
+    @staticmethod
+    def state_dict_converter():
+        return WanControlNetModelStateDictConverter()
+    
+    
+class WanControlNetModelStateDictConverter:
+    def __init__(self):
+        pass
+    
+    def from_diffusers(self, state_dict):
+        return state_dict
+    
+    def from_civitai(self, state_dict):
+        return state_dict
+    
+    def from_base_model(self, state_dict):
+        if hash_state_dict_keys(state_dict) == "9269f8db9040a9d860eaca435be61814":
+            config = {
+                "has_image_input": False,
+                "patch_size": [1, 2, 2],
+                "in_dim": 16,
+                "dim": 1536,
+                "ffn_dim": 8960,
+                "freq_dim": 256,
+                "text_dim": 4096,
+                "out_dim": 16,
+                "num_heads": 12,
+                "num_layers": 30,
+                "eps": 1e-6
+            }
+        elif hash_state_dict_keys(state_dict) == "aafcfd9672c3a2456dc46e1cb6e52c70":
+            config = {
+                "has_image_input": False,
+                "patch_size": [1, 2, 2],
+                "in_dim": 16,
+                "dim": 5120,
+                "ffn_dim": 13824,
+                "freq_dim": 256,
+                "text_dim": 4096,
+                "out_dim": 16,
+                "num_heads": 40,
+                "num_layers": 40,
+                "eps": 1e-6
+            }
+        elif hash_state_dict_keys(state_dict) == "6bfcfb3b342cb286ce886889d519a77e":
+            config = {
+                "has_image_input": True,
+                "patch_size": [1, 2, 2],
+                "in_dim": 36,
+                "dim": 5120,
+                "ffn_dim": 13824,
+                "freq_dim": 256,
+                "text_dim": 4096,
+                "out_dim": 16,
+                "num_heads": 40,
+                "num_layers": 40,
+                "eps": 1e-6
+            }
+        else:
+            config = {}
+        state_dict_ = {}
+        dtype, device = None, None
+        for name, param in state_dict.items():
+            if name.startswith("head."):
+                continue
+            state_dict_[name] = param
+            dtype, device = param.dtype, param.device
+        for block_id in range(config["num_layers"]):
+            zeros = torch.zeros((config["dim"], config["dim"]), dtype=dtype, device=device)
+            state_dict_[f"controlnet_blocks.{block_id}.weight"] = zeros.clone()
+        state_dict_["controlnet_conv_in.weight"] = torch.zeros((config["in_dim"], config["in_dim"], 1, 1, 1), dtype=dtype, device=device)
+        state_dict_["controlnet_conv_in.bias"] = torch.zeros((config["in_dim"],), dtype=dtype, device=device)
+        return state_dict_, config

diffsynth/models/wan_video_motion_controller.py

@@ -0,0 +1,27 @@
+import torch
+import torch.nn as nn
+from .wan_video_dit import sinusoidal_embedding_1d
+
+
+
+class WanMotionControllerModel(torch.nn.Module):
+    def __init__(self, freq_dim=256, dim=1536):
+        super().__init__()
+        self.freq_dim = freq_dim
+        self.linear = nn.Sequential(
+            nn.Linear(freq_dim, dim),
+            nn.SiLU(),
+            nn.Linear(dim, dim),
+            nn.SiLU(),
+            nn.Linear(dim, dim * 6),
+        )
+        
+    def forward(self, motion_bucket_id):
+        emb = sinusoidal_embedding_1d(self.freq_dim, motion_bucket_id * 10)
+        emb = self.linear(emb)
+        return emb
+    
+    def init(self):
+        state_dict = self.linear[-1].state_dict()
+        state_dict = {i: state_dict[i] * 0 for i in state_dict}
+        self.linear[-1].load_state_dict(state_dict)

diffsynth/pipelines/wan_video.py

@@ -17,6 +17,8 @@ from ..vram_management import enable_vram_management, AutoWrappedModule, AutoWra
 from ..models.wan_video_text_encoder import T5RelativeEmbedding, T5LayerNorm
 from ..models.wan_video_dit import RMSNorm, sinusoidal_embedding_1d
 from ..models.wan_video_vae import RMS_norm, CausalConv3d, Upsample
+from ..models.wan_video_controlnet import WanControlNetModel
+from ..models.wan_video_motion_controller import WanMotionControllerModel
 
 
 
@@ -30,7 +32,9 @@ class WanVideoPipeline(BasePipeline):
         self.image_encoder: WanImageEncoder = None
         self.dit: WanModel = None
         self.vae: WanVideoVAE = None
-        self.model_names = ['text_encoder', 'dit', 'vae']
+        self.controlnet: WanControlNetModel = None
+        self.motion_controller: WanMotionControllerModel = None
+        self.model_names = ['text_encoder', 'dit', 'vae', 'image_encoder', 'controlnet', 'motion_controller']
         self.height_division_factor = 16
         self.width_division_factor = 16
 
@@ -189,6 +193,16 @@ class WanVideoPipeline(BasePipeline):
     def decode_video(self, latents, tiled=True, tile_size=(34, 34), tile_stride=(18, 16)):
         frames = self.vae.decode(latents, device=self.device, tiled=tiled, tile_size=tile_size, tile_stride=tile_stride)
         return frames
+    
+    
+    def prepare_controlnet(self, controlnet_frames, tiled=True, tile_size=(34, 34), tile_stride=(18, 16)):
+        controlnet_conditioning = self.encode_video(controlnet_frames, tiled=tiled, tile_size=tile_size, tile_stride=tile_stride).to(dtype=self.torch_dtype, device=self.device)
+        return {"controlnet_conditioning": controlnet_conditioning}
+    
+    
+    def prepare_motion_bucket_id(self, motion_bucket_id):
+        motion_bucket_id = torch.Tensor((motion_bucket_id,)).to(dtype=self.torch_dtype, device=self.device)
+        return {"motion_bucket_id": motion_bucket_id}
 
 
     @torch.no_grad()
@@ -207,11 +221,13 @@ class WanVideoPipeline(BasePipeline):
         cfg_scale=5.0,
         num_inference_steps=50,
         sigma_shift=5.0,
+        motion_bucket_id=None,
         tiled=True,
         tile_size=(30, 52),
         tile_stride=(15, 26),
         tea_cache_l1_thresh=None,
         tea_cache_model_id="",
+        controlnet_frames=None,
         progress_bar_cmd=tqdm,
         progress_bar_st=None,
     ):
@@ -252,6 +268,21 @@ class WanVideoPipeline(BasePipeline):
         else:
             image_emb = {}
             
+        # ControlNet
+        if self.controlnet is not None and controlnet_frames is not None:
+            self.load_models_to_device(['vae', 'controlnet'])
+            controlnet_frames = self.preprocess_images(controlnet_frames)
+            controlnet_frames = torch.stack(controlnet_frames, dim=2).to(dtype=self.torch_dtype, device=self.device)
+            controlnet_kwargs = self.prepare_controlnet(controlnet_frames)
+        else:
+            controlnet_kwargs = {}
+            
+        # Motion Controller
+        if self.motion_controller is not None and motion_bucket_id is not None:
+            motion_kwargs = self.prepare_motion_bucket_id(motion_bucket_id)
+        else:
+            motion_kwargs = {}
+            
         # Extra input
         extra_input = self.prepare_extra_input(latents)
         
@@ -260,14 +291,24 @@ class WanVideoPipeline(BasePipeline):
         tea_cache_nega = {"tea_cache": TeaCache(num_inference_steps, rel_l1_thresh=tea_cache_l1_thresh, model_id=tea_cache_model_id) if tea_cache_l1_thresh is not None else None}
 
         # Denoise
-        self.load_models_to_device(["dit"])
+        self.load_models_to_device(["dit", "controlnet", "motion_controller"])
         for progress_id, timestep in enumerate(progress_bar_cmd(self.scheduler.timesteps)):
             timestep = timestep.unsqueeze(0).to(dtype=self.torch_dtype, device=self.device)
 
             # Inference
-            noise_pred_posi = model_fn_wan_video(self.dit, latents, timestep=timestep, **prompt_emb_posi, **image_emb, **extra_input, **tea_cache_posi)
+            noise_pred_posi = model_fn_wan_video(
+                self.dit, controlnet=self.controlnet, motion_controller=self.motion_controller,
+                x=latents, timestep=timestep,
+                **prompt_emb_posi, **image_emb, **extra_input,
+                **tea_cache_posi, **controlnet_kwargs, **motion_kwargs,
+            )
             if cfg_scale != 1.0:
-                noise_pred_nega = model_fn_wan_video(self.dit, latents, timestep=timestep, **prompt_emb_nega, **image_emb, **extra_input, **tea_cache_nega)
+                noise_pred_nega = model_fn_wan_video(
+                    self.dit, controlnet=self.controlnet, motion_controller=self.motion_controller,
+                    x=latents, timestep=timestep,
+                    **prompt_emb_nega, **image_emb, **extra_input,
+                    **tea_cache_nega, **controlnet_kwargs, **motion_kwargs,
+                )
                 noise_pred = noise_pred_nega + cfg_scale * (noise_pred_posi - noise_pred_nega)
             else:
                 noise_pred = noise_pred_posi
@@ -340,16 +381,35 @@ class TeaCache:
 
 def model_fn_wan_video(
     dit: WanModel,
-    x: torch.Tensor,
-    timestep: torch.Tensor,
-    context: torch.Tensor,
+    controlnet: WanControlNetModel = None,
+    motion_controller: WanMotionControllerModel = None,
+    x: torch.Tensor = None,
+    timestep: torch.Tensor = None,
+    context: torch.Tensor = None,
     clip_feature: Optional[torch.Tensor] = None,
     y: Optional[torch.Tensor] = None,
     tea_cache: TeaCache = None,
+    controlnet_conditioning: Optional[torch.Tensor] = None,
+    motion_bucket_id: Optional[torch.Tensor] = None,
+    use_gradient_checkpointing: bool = False,
+    use_gradient_checkpointing_offload: bool = False,
     **kwargs,
 ):
+    # ControlNet
+    if controlnet is not None and controlnet_conditioning is not None:
+        controlnet_res_stack = controlnet(
+            x, timestep=timestep, context=context, clip_feature=clip_feature, y=y,
+            controlnet_conditioning=controlnet_conditioning,
+            use_gradient_checkpointing=use_gradient_checkpointing,
+            use_gradient_checkpointing_offload=use_gradient_checkpointing_offload,
+        )
+    else:
+        controlnet_res_stack = None
+    
     t = dit.time_embedding(sinusoidal_embedding_1d(dit.freq_dim, timestep))
     t_mod = dit.time_projection(t).unflatten(1, (6, dit.dim))
+    if motion_bucket_id is not None and motion_controller is not None:
+        t_mod = t_mod + motion_controller(motion_bucket_id).unflatten(1, (6, dit.dim))
     context = dit.text_embedding(context)
     
     if dit.has_image_input:
@@ -370,13 +430,35 @@ def model_fn_wan_video(
         tea_cache_update = tea_cache.check(dit, x, t_mod)
     else:
         tea_cache_update = False
+        
+    def create_custom_forward(module):
+        def custom_forward(*inputs):
+            return module(*inputs)
+        return custom_forward
     
     if tea_cache_update:
         x = tea_cache.update(x)
     else:
         # blocks
-        for block in dit.blocks:
-            x = block(x, context, t_mod, freqs)
+        for block_id, block in enumerate(dit.blocks):
+            if dit.training and use_gradient_checkpointing:
+                if use_gradient_checkpointing_offload:
+                    with torch.autograd.graph.save_on_cpu():
+                        x = torch.utils.checkpoint.checkpoint(
+                            create_custom_forward(block),
+                            x, context, t_mod, freqs,
+                            use_reentrant=False,
+                        )
+                else:
+                    x = torch.utils.checkpoint.checkpoint(
+                        create_custom_forward(block),
+                        x, context, t_mod, freqs,
+                        use_reentrant=False,
+                    )
+            else:
+                x = block(x, context, t_mod, freqs)
+            if controlnet_res_stack is not None:
+                x = x + controlnet_res_stack[block_id]
         if tea_cache is not None:
             tea_cache.store(x)
 

examples/wanvideo/train_wan_t2v.py

@@ -12,9 +12,12 @@ import numpy as np
 
 
 class TextVideoDataset(torch.utils.data.Dataset):
-    def __init__(self, base_path, metadata_path, max_num_frames=81, frame_interval=1, num_frames=81, height=480, width=832, is_i2v=False):
+    def __init__(self, base_path, metadata_path, max_num_frames=81, frame_interval=1, num_frames=81, height=480, width=832, is_i2v=False, target_fps=None):
         metadata = pd.read_csv(metadata_path)
-        self.path = [os.path.join(base_path, "train", file_name) for file_name in metadata["file_name"]]
+        if os.path.exists(os.path.join(base_path, "train")):
+            self.path = [os.path.join(base_path, "train", file_name) for file_name in metadata["file_name"]]
+        else:
+            self.path = [os.path.join(base_path, file_name) for file_name in metadata["file_name"]]
         self.text = metadata["text"].to_list()
         
         self.max_num_frames = max_num_frames
@@ -23,6 +26,7 @@ class TextVideoDataset(torch.utils.data.Dataset):
         self.height = height
         self.width = width
         self.is_i2v = is_i2v
+        self.target_fps = target_fps
             
         self.frame_process = v2.Compose([
             v2.CenterCrop(size=(height, width)),
@@ -71,8 +75,15 @@ class TextVideoDataset(torch.utils.data.Dataset):
 
 
     def load_video(self, file_path):
-        start_frame_id = torch.randint(0, self.max_num_frames - (self.num_frames - 1) * self.frame_interval, (1,))[0]
-        frames = self.load_frames_using_imageio(file_path, self.max_num_frames, start_frame_id, self.frame_interval, self.num_frames, self.frame_process)
+        start_frame_id = 0
+        if self.target_fps is None:
+            frame_interval = self.frame_interval
+        else:
+            reader = imageio.get_reader(file_path)
+            fps = reader.get_meta_data()["fps"]
+            reader.close()
+            frame_interval = max(round(fps / self.target_fps), 1)
+        frames = self.load_frames_using_imageio(file_path, self.max_num_frames, start_frame_id, frame_interval, self.num_frames, self.frame_process)
         return frames
     
     
@@ -95,17 +106,20 @@ class TextVideoDataset(torch.utils.data.Dataset):
     def __getitem__(self, data_id):
         text = self.text[data_id]
         path = self.path[data_id]
-        if self.is_image(path):
+        try:
+            if self.is_image(path):
+                if self.is_i2v:
+                    raise ValueError(f"{path} is not a video. I2V model doesn't support image-to-image training.")
+                video = self.load_image(path)
+            else:
+                video = self.load_video(path)
             if self.is_i2v:
-                raise ValueError(f"{path} is not a video. I2V model doesn't support image-to-image training.")
-            video = self.load_image(path)
-        else:
-            video = self.load_video(path)
-        if self.is_i2v:
-            video, first_frame = video
-            data = {"text": text, "video": video, "path": path, "first_frame": first_frame}
-        else:
-            data = {"text": text, "video": video, "path": path}
+                video, first_frame = video
+                data = {"text": text, "video": video, "path": path, "first_frame": first_frame}
+            else:
+                data = {"text": text, "video": video, "path": path}
+        except:
+            data = None
         return data
     
 
@@ -115,7 +129,7 @@ class TextVideoDataset(torch.utils.data.Dataset):
 
 
 class LightningModelForDataProcess(pl.LightningModule):
-    def __init__(self, text_encoder_path, vae_path, image_encoder_path=None, tiled=False, tile_size=(34, 34), tile_stride=(18, 16)):
+    def __init__(self, text_encoder_path, vae_path, image_encoder_path=None, tiled=False, tile_size=(34, 34), tile_stride=(18, 16), redirected_tensor_path=None):
         super().__init__()
         model_path = [text_encoder_path, vae_path]
         if image_encoder_path is not None:
@@ -125,9 +139,13 @@ class LightningModelForDataProcess(pl.LightningModule):
         self.pipe = WanVideoPipeline.from_model_manager(model_manager)
 
         self.tiler_kwargs = {"tiled": tiled, "tile_size": tile_size, "tile_stride": tile_stride}
+        self.redirected_tensor_path = redirected_tensor_path
         
     def test_step(self, batch, batch_idx):
-        text, video, path = batch["text"][0], batch["video"], batch["path"][0]
+        data = batch[0]
+        if data is None or data["video"] is None:
+            return
+        text, video, path = data["text"], data["video"].unsqueeze(0), data["path"]
         
         self.pipe.device = self.device
         if video is not None:
@@ -144,28 +162,49 @@ class LightningModelForDataProcess(pl.LightningModule):
             else:
                 image_emb = {}
             data = {"latents": latents, "prompt_emb": prompt_emb, "image_emb": image_emb}
+            if self.redirected_tensor_path is not None:
+                path = path.replace("/", "_").replace("\\", "_")
+                path = os.path.join(self.redirected_tensor_path, path)
             torch.save(data, path + ".tensors.pth")
 
 
 
 class TensorDataset(torch.utils.data.Dataset):
-    def __init__(self, base_path, metadata_path, steps_per_epoch):
-        metadata = pd.read_csv(metadata_path)
-        self.path = [os.path.join(base_path, "train", file_name) for file_name in metadata["file_name"]]
-        print(len(self.path), "videos in metadata.")
-        self.path = [i + ".tensors.pth" for i in self.path if os.path.exists(i + ".tensors.pth")]
+    def __init__(self, base_path, metadata_path=None, steps_per_epoch=1000, redirected_tensor_path=None):
+        if os.path.exists(metadata_path):
+            metadata = pd.read_csv(metadata_path)
+            self.path = [os.path.join(base_path, "train", file_name) for file_name in metadata["file_name"]]
+            print(len(self.path), "videos in metadata.")
+            if redirected_tensor_path is None:
+                self.path = [i + ".tensors.pth" for i in self.path if os.path.exists(i + ".tensors.pth")]
+            else:
+                cached_path = []
+                for path in self.path:
+                    path = path.replace("/", "_").replace("\\", "_")
+                    path = os.path.join(redirected_tensor_path, path)
+                    if os.path.exists(path + ".tensors.pth"):
+                        cached_path.append(path + ".tensors.pth")
+                self.path = cached_path
+        else:
+            print("Cannot find metadata.csv. Trying to search for tensor files.")
+            self.path = [os.path.join(base_path, i) for i in os.listdir(base_path) if i.endswith(".tensors.pth")]
         print(len(self.path), "tensors cached in metadata.")
         assert len(self.path) > 0
         
         self.steps_per_epoch = steps_per_epoch
+        self.redirected_tensor_path = redirected_tensor_path
 
 
     def __getitem__(self, index):
-        data_id = torch.randint(0, len(self.path), (1,))[0]
-        data_id = (data_id + index) % len(self.path) # For fixed seed.
-        path = self.path[data_id]
-        data = torch.load(path, weights_only=True, map_location="cpu")
-        return data
+        while True:
+            try:
+                data_id = torch.randint(0, len(self.path), (1,))[0]
+                data_id = (data_id + index) % len(self.path) # For fixed seed.
+                path = self.path[data_id]
+                data = torch.load(path, weights_only=True, map_location="cpu")
+                return data
+            except:
+                continue
     
 
     def __len__(self):
@@ -323,6 +362,18 @@ def parse_args():
         default="./",
         help="Path to save the model.",
     )
+    parser.add_argument(
+        "--metadata_path",
+        type=str,
+        default=None,
+        help="Path to metadata.csv.",
+    )
+    parser.add_argument(
+        "--redirected_tensor_path",
+        type=str,
+        default=None,
+        help="Path to save cached tensors.",
+    )
     parser.add_argument(
         "--text_encoder_path",
         type=str,
@@ -389,6 +440,12 @@ def parse_args():
         default=81,
         help="Number of frames.",
     )
+    parser.add_argument(
+        "--target_fps",
+        type=int,
+        default=None,
+        help="Expected FPS for sampling frames.",
+    )
     parser.add_argument(
         "--height",
         type=int,
@@ -500,19 +557,21 @@ def parse_args():
 def data_process(args):
     dataset = TextVideoDataset(
         args.dataset_path,
-        os.path.join(args.dataset_path, "metadata.csv"),
+        os.path.join(args.dataset_path, "metadata.csv") if args.metadata_path is None else args.metadata_path,
         max_num_frames=args.num_frames,
         frame_interval=1,
         num_frames=args.num_frames,
         height=args.height,
         width=args.width,
-        is_i2v=args.image_encoder_path is not None
+        is_i2v=args.image_encoder_path is not None,
+        target_fps=args.target_fps,
     )
     dataloader = torch.utils.data.DataLoader(
         dataset,
         shuffle=False,
         batch_size=1,
-        num_workers=args.dataloader_num_workers
+        num_workers=args.dataloader_num_workers,
+        collate_fn=lambda x: x,
     )
     model = LightningModelForDataProcess(
         text_encoder_path=args.text_encoder_path,
@@ -521,6 +580,7 @@ def data_process(args):
         tiled=args.tiled,
         tile_size=(args.tile_size_height, args.tile_size_width),
         tile_stride=(args.tile_stride_height, args.tile_stride_width),
+        redirected_tensor_path=args.redirected_tensor_path,
     )
     trainer = pl.Trainer(
         accelerator="gpu",
@@ -533,8 +593,9 @@ def data_process(args):
 def train(args):
     dataset = TensorDataset(
         args.dataset_path,
-        os.path.join(args.dataset_path, "metadata.csv"),
+        os.path.join(args.dataset_path, "metadata.csv") if args.metadata_path is None else args.metadata_path,
         steps_per_epoch=args.steps_per_epoch,
+        redirected_tensor_path=args.redirected_tensor_path,
     )
     dataloader = torch.utils.data.DataLoader(
         dataset,

examples/wanvideo/train_wan_t2v_controlnet.py

@@ -0,0 +1,626 @@
+import torch, os, imageio, argparse
+from torchvision.transforms import v2
+from einops import rearrange
+import lightning as pl
+import pandas as pd
+from diffsynth import WanVideoPipeline, ModelManager, load_state_dict
+from peft import LoraConfig, inject_adapter_in_model
+import torchvision
+from PIL import Image
+import numpy as np
+from diffsynth.models.wan_video_controlnet import WanControlNetModel
+from diffsynth.pipelines.wan_video import model_fn_wan_video
+
+
+
+class TextVideoDataset(torch.utils.data.Dataset):
+    def __init__(self, base_path, metadata_path, max_num_frames=81, frame_interval=1, num_frames=81, height=480, width=832, is_i2v=False, target_fps=None):
+        metadata = pd.read_csv(metadata_path)
+        self.path = [os.path.join(base_path, "train", file_name) for file_name in metadata["file_name"]]
+        self.controlnet_path = [os.path.join(base_path, file_name) for file_name in metadata["controlnet_file_name"]]
+        self.text = metadata["text"].to_list()
+        
+        self.max_num_frames = max_num_frames
+        self.frame_interval = frame_interval
+        self.num_frames = num_frames
+        self.height = height
+        self.width = width
+        self.is_i2v = is_i2v
+        self.target_fps = target_fps
+            
+        self.frame_process = v2.Compose([
+            v2.CenterCrop(size=(height, width)),
+            v2.Resize(size=(height, width), antialias=True),
+            v2.ToTensor(),
+            v2.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5]),
+        ])
+        
+        
+    def crop_and_resize(self, image):
+        width, height = image.size
+        scale = max(self.width / width, self.height / height)
+        image = torchvision.transforms.functional.resize(
+            image,
+            (round(height*scale), round(width*scale)),
+            interpolation=torchvision.transforms.InterpolationMode.BILINEAR
+        )
+        return image
+
+
+    def load_frames_using_imageio(self, file_path, max_num_frames, start_frame_id, interval, num_frames, frame_process):
+        reader = imageio.get_reader(file_path)
+        if reader.count_frames() < max_num_frames or reader.count_frames() - 1 < start_frame_id + (num_frames - 1) * interval:
+            reader.close()
+            return None
+        
+        frames = []
+        first_frame = None
+        for frame_id in range(num_frames):
+            frame = reader.get_data(start_frame_id + frame_id * interval)
+            frame = Image.fromarray(frame)
+            frame = self.crop_and_resize(frame)
+            if first_frame is None:
+                first_frame = np.array(frame)
+            frame = frame_process(frame)
+            frames.append(frame)
+        reader.close()
+
+        frames = torch.stack(frames, dim=0)
+        frames = rearrange(frames, "T C H W -> C T H W")
+
+        if self.is_i2v:
+            return frames, first_frame
+        else:
+            return frames
+
+
+    def load_video(self, file_path):
+        start_frame_id = 0
+        if self.target_fps is None:
+            frame_interval = self.frame_interval
+        else:
+            reader = imageio.get_reader(file_path)
+            fps = reader.get_meta_data()["fps"]
+            reader.close()
+            frame_interval = max(round(fps / self.target_fps), 1)
+        frames = self.load_frames_using_imageio(file_path, self.max_num_frames, start_frame_id, frame_interval, self.num_frames, self.frame_process)
+        return frames
+    
+    
+    def is_image(self, file_path):
+        file_ext_name = file_path.split(".")[-1]
+        if file_ext_name.lower() in ["jpg", "jpeg", "png", "webp"]:
+            return True
+        return False
+    
+    
+    def load_image(self, file_path):
+        frame = Image.open(file_path).convert("RGB")
+        frame = self.crop_and_resize(frame)
+        frame = self.frame_process(frame)
+        frame = rearrange(frame, "C H W -> C 1 H W")
+        return frame
+
+
+    def __getitem__(self, data_id):
+        text = self.text[data_id]
+        path = self.path[data_id]
+        controlnet_path = self.controlnet_path[data_id]
+        try:
+            if self.is_image(path):
+                if self.is_i2v:
+                    raise ValueError(f"{path} is not a video. I2V model doesn't support image-to-image training.")
+                video = self.load_image(path)
+            else:
+                video = self.load_video(path)
+                controlnet_frames = self.load_video(controlnet_path)
+            if self.is_i2v:
+                video, first_frame = video
+                data = {"text": text, "video": video, "path": path, "first_frame": first_frame}
+            else:
+                data = {"text": text, "video": video, "path": path, "controlnet_frames": controlnet_frames}
+        except:
+            data = None
+        return data
+    
+
+    def __len__(self):
+        return len(self.path)
+
+
+
+class LightningModelForDataProcess(pl.LightningModule):
+    def __init__(self, text_encoder_path, vae_path, image_encoder_path=None, tiled=False, tile_size=(34, 34), tile_stride=(18, 16), redirected_tensor_path=None):
+        super().__init__()
+        model_path = [text_encoder_path, vae_path]
+        if image_encoder_path is not None:
+            model_path.append(image_encoder_path)
+        model_manager = ModelManager(torch_dtype=torch.bfloat16, device="cpu")
+        model_manager.load_models(model_path)
+        self.pipe = WanVideoPipeline.from_model_manager(model_manager)
+
+        self.tiler_kwargs = {"tiled": tiled, "tile_size": tile_size, "tile_stride": tile_stride}
+        self.redirected_tensor_path = redirected_tensor_path
+        
+    def test_step(self, batch, batch_idx):
+        data = batch[0]
+        if data is None or data["video"] is None:
+            return
+        text, video, path = data["text"], data["video"].unsqueeze(0), data["path"]
+        controlnet_frames = data["controlnet_frames"].unsqueeze(0)
+        
+        self.pipe.device = self.device
+        if video is not None:
+            # prompt
+            prompt_emb = self.pipe.encode_prompt(text)
+            # video
+            video = video.to(dtype=self.pipe.torch_dtype, device=self.pipe.device)
+            latents = self.pipe.encode_video(video, **self.tiler_kwargs)[0]
+            # ControlNet video
+            controlnet_frames = controlnet_frames.to(dtype=self.pipe.torch_dtype, device=self.pipe.device)
+            controlnet_kwargs = self.pipe.prepare_controlnet(controlnet_frames, **self.tiler_kwargs)
+            controlnet_kwargs["controlnet_conditioning"] = controlnet_kwargs["controlnet_conditioning"][0]
+            # image
+            if "first_frame" in batch:
+                first_frame = Image.fromarray(batch["first_frame"][0].cpu().numpy())
+                _, _, num_frames, height, width = video.shape
+                image_emb = self.pipe.encode_image(first_frame, num_frames, height, width)
+            else:
+                image_emb = {}
+            data = {"latents": latents, "prompt_emb": prompt_emb, "image_emb": image_emb, "controlnet_kwargs": controlnet_kwargs}
+            if self.redirected_tensor_path is not None:
+                path = path.replace("/", "_").replace("\\", "_")
+                path = os.path.join(self.redirected_tensor_path, path)
+            torch.save(data, path + ".tensors.pth")
+
+
+
+class TensorDataset(torch.utils.data.Dataset):
+    def __init__(self, base_path, metadata_path=None, steps_per_epoch=1000, redirected_tensor_path=None):
+        if os.path.exists(metadata_path):
+            metadata = pd.read_csv(metadata_path)
+            self.path = [os.path.join(base_path, "train", file_name) for file_name in metadata["file_name"]]
+            print(len(self.path), "videos in metadata.")
+            if redirected_tensor_path is None:
+                self.path = [i + ".tensors.pth" for i in self.path if os.path.exists(i + ".tensors.pth")]
+            else:
+                cached_path = []
+                for path in self.path:
+                    path = path.replace("/", "_").replace("\\", "_")
+                    path = os.path.join(redirected_tensor_path, path)
+                    if os.path.exists(path + ".tensors.pth"):
+                        cached_path.append(path + ".tensors.pth")
+                self.path = cached_path
+        else:
+            print("Cannot find metadata.csv. Trying to search for tensor files.")
+            self.path = [os.path.join(base_path, i) for i in os.listdir(base_path) if i.endswith(".tensors.pth")]
+        print(len(self.path), "tensors cached in metadata.")
+        assert len(self.path) > 0
+        
+        self.steps_per_epoch = steps_per_epoch
+        self.redirected_tensor_path = redirected_tensor_path
+
+
+    def __getitem__(self, index):
+        while True:
+            try:
+                data_id = torch.randint(0, len(self.path), (1,))[0]
+                data_id = (data_id + index) % len(self.path) # For fixed seed.
+                path = self.path[data_id]
+                data = torch.load(path, weights_only=True, map_location="cpu")
+                return data
+            except:
+                continue
+    
+
+    def __len__(self):
+        return self.steps_per_epoch
+
+
+
+class LightningModelForTrain(pl.LightningModule):
+    def __init__(
+        self,
+        dit_path,
+        learning_rate=1e-5,
+        lora_rank=4, lora_alpha=4, train_architecture="lora", lora_target_modules="q,k,v,o,ffn.0,ffn.2", init_lora_weights="kaiming",
+        use_gradient_checkpointing=True, use_gradient_checkpointing_offload=False,
+        pretrained_lora_path=None
+    ):
+        super().__init__()
+        model_manager = ModelManager(torch_dtype=torch.bfloat16, device="cpu")
+        if os.path.isfile(dit_path):
+            model_manager.load_models([dit_path])
+        else:
+            dit_path = dit_path.split(",")
+            model_manager.load_models([dit_path])
+        
+        self.pipe = WanVideoPipeline.from_model_manager(model_manager)
+        self.pipe.scheduler.set_timesteps(1000, training=True)
+        self.freeze_parameters()
+        
+        state_dict = load_state_dict(dit_path, torch_dtype=torch.bfloat16)
+        state_dict, config = WanControlNetModel.state_dict_converter().from_base_model(state_dict)
+        self.pipe.controlnet = WanControlNetModel(**config).to(torch.bfloat16)
+        self.pipe.controlnet.load_state_dict(state_dict)
+        self.pipe.controlnet.train()
+        self.pipe.controlnet.requires_grad_(True)
+        
+        self.learning_rate = learning_rate
+        self.use_gradient_checkpointing = use_gradient_checkpointing
+        self.use_gradient_checkpointing_offload = use_gradient_checkpointing_offload
+        
+        
+    def freeze_parameters(self):
+        # Freeze parameters
+        self.pipe.requires_grad_(False)
+        self.pipe.eval()
+        self.pipe.denoising_model().train()
+
+
+    def training_step(self, batch, batch_idx):
+        # Data
+        latents = batch["latents"].to(self.device)
+        controlnet_kwargs = batch["controlnet_kwargs"]
+        controlnet_kwargs["controlnet_conditioning"] = controlnet_kwargs["controlnet_conditioning"].to(self.device)
+        prompt_emb = batch["prompt_emb"]
+        prompt_emb["context"] = prompt_emb["context"][0].to(self.device)
+        image_emb = batch["image_emb"]
+        if "clip_feature" in image_emb:
+            image_emb["clip_feature"] = image_emb["clip_feature"][0].to(self.device)
+        if "y" in image_emb:
+            image_emb["y"] = image_emb["y"][0].to(self.device)
+
+        # Loss
+        self.pipe.device = self.device
+        noise = torch.randn_like(latents)
+        timestep_id = torch.randint(0, self.pipe.scheduler.num_train_timesteps, (1,))
+        timestep = self.pipe.scheduler.timesteps[timestep_id].to(dtype=self.pipe.torch_dtype, device=self.pipe.device)
+        extra_input = self.pipe.prepare_extra_input(latents)
+        noisy_latents = self.pipe.scheduler.add_noise(latents, noise, timestep)
+        training_target = self.pipe.scheduler.training_target(latents, noise, timestep)
+
+        # Compute loss
+        noise_pred = model_fn_wan_video(
+            dit=self.pipe.dit, controlnet=self.pipe.controlnet,
+            x=noisy_latents, timestep=timestep, **prompt_emb, **extra_input, **image_emb, **controlnet_kwargs,
+            use_gradient_checkpointing=self.use_gradient_checkpointing,
+            use_gradient_checkpointing_offload=self.use_gradient_checkpointing_offload
+        )
+        loss = torch.nn.functional.mse_loss(noise_pred.float(), training_target.float())
+        loss = loss * self.pipe.scheduler.training_weight(timestep)
+
+        # Record log
+        self.log("train_loss", loss, prog_bar=True)
+        return loss
+
+
+    def configure_optimizers(self):
+        trainable_modules = filter(lambda p: p.requires_grad, self.pipe.controlnet.parameters())
+        optimizer = torch.optim.AdamW(trainable_modules, lr=self.learning_rate)
+        return optimizer
+    
+
+    def on_save_checkpoint(self, checkpoint):
+        checkpoint.clear()
+        trainable_param_names = list(filter(lambda named_param: named_param[1].requires_grad, self.pipe.controlnet.named_parameters()))
+        trainable_param_names = set([named_param[0] for named_param in trainable_param_names])
+        state_dict = self.pipe.controlnet.state_dict()
+        lora_state_dict = {}
+        for name, param in state_dict.items():
+            if name in trainable_param_names:
+                lora_state_dict[name] = param
+        checkpoint.update(lora_state_dict)
+
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(description="Simple example of a training script.")
+    parser.add_argument(
+        "--task",
+        type=str,
+        default="data_process",
+        required=True,
+        choices=["data_process", "train"],
+        help="Task. `data_process` or `train`.",
+    )
+    parser.add_argument(
+        "--dataset_path",
+        type=str,
+        default=None,
+        required=True,
+        help="The path of the Dataset.",
+    )
+    parser.add_argument(
+        "--output_path",
+        type=str,
+        default="./",
+        help="Path to save the model.",
+    )
+    parser.add_argument(
+        "--metadata_path",
+        type=str,
+        default=None,
+        help="Path to metadata.csv.",
+    )
+    parser.add_argument(
+        "--redirected_tensor_path",
+        type=str,
+        default=None,
+        help="Path to save cached tensors.",
+    )
+    parser.add_argument(
+        "--text_encoder_path",
+        type=str,
+        default=None,
+        help="Path of text encoder.",
+    )
+    parser.add_argument(
+        "--image_encoder_path",
+        type=str,
+        default=None,
+        help="Path of image encoder.",
+    )
+    parser.add_argument(
+        "--vae_path",
+        type=str,
+        default=None,
+        help="Path of VAE.",
+    )
+    parser.add_argument(
+        "--dit_path",
+        type=str,
+        default=None,
+        help="Path of DiT.",
+    )
+    parser.add_argument(
+        "--tiled",
+        default=False,
+        action="store_true",
+        help="Whether enable tile encode in VAE. This option can reduce VRAM required.",
+    )
+    parser.add_argument(
+        "--tile_size_height",
+        type=int,
+        default=34,
+        help="Tile size (height) in VAE.",
+    )
+    parser.add_argument(
+        "--tile_size_width",
+        type=int,
+        default=34,
+        help="Tile size (width) in VAE.",
+    )
+    parser.add_argument(
+        "--tile_stride_height",
+        type=int,
+        default=18,
+        help="Tile stride (height) in VAE.",
+    )
+    parser.add_argument(
+        "--tile_stride_width",
+        type=int,
+        default=16,
+        help="Tile stride (width) in VAE.",
+    )
+    parser.add_argument(
+        "--steps_per_epoch",
+        type=int,
+        default=500,
+        help="Number of steps per epoch.",
+    )
+    parser.add_argument(
+        "--num_frames",
+        type=int,
+        default=81,
+        help="Number of frames.",
+    )
+    parser.add_argument(
+        "--target_fps",
+        type=int,
+        default=None,
+        help="Expected FPS for sampling frames.",
+    )
+    parser.add_argument(
+        "--height",
+        type=int,
+        default=480,
+        help="Image height.",
+    )
+    parser.add_argument(
+        "--width",
+        type=int,
+        default=832,
+        help="Image width.",
+    )
+    parser.add_argument(
+        "--dataloader_num_workers",
+        type=int,
+        default=1,
+        help="Number of subprocesses to use for data loading. 0 means that the data will be loaded in the main process.",
+    )
+    parser.add_argument(
+        "--learning_rate",
+        type=float,
+        default=1e-5,
+        help="Learning rate.",
+    )
+    parser.add_argument(
+        "--accumulate_grad_batches",
+        type=int,
+        default=1,
+        help="The number of batches in gradient accumulation.",
+    )
+    parser.add_argument(
+        "--max_epochs",
+        type=int,
+        default=1,
+        help="Number of epochs.",
+    )
+    parser.add_argument(
+        "--lora_target_modules",
+        type=str,
+        default="q,k,v,o,ffn.0,ffn.2",
+        help="Layers with LoRA modules.",
+    )
+    parser.add_argument(
+        "--init_lora_weights",
+        type=str,
+        default="kaiming",
+        choices=["gaussian", "kaiming"],
+        help="The initializing method of LoRA weight.",
+    )
+    parser.add_argument(
+        "--training_strategy",
+        type=str,
+        default="auto",
+        choices=["auto", "deepspeed_stage_1", "deepspeed_stage_2", "deepspeed_stage_3"],
+        help="Training strategy",
+    )
+    parser.add_argument(
+        "--lora_rank",
+        type=int,
+        default=4,
+        help="The dimension of the LoRA update matrices.",
+    )
+    parser.add_argument(
+        "--lora_alpha",
+        type=float,
+        default=4.0,
+        help="The weight of the LoRA update matrices.",
+    )
+    parser.add_argument(
+        "--use_gradient_checkpointing",
+        default=False,
+        action="store_true",
+        help="Whether to use gradient checkpointing.",
+    )
+    parser.add_argument(
+        "--use_gradient_checkpointing_offload",
+        default=False,
+        action="store_true",
+        help="Whether to use gradient checkpointing offload.",
+    )
+    parser.add_argument(
+        "--train_architecture",
+        type=str,
+        default="lora",
+        choices=["lora", "full"],
+        help="Model structure to train. LoRA training or full training.",
+    )
+    parser.add_argument(
+        "--pretrained_lora_path",
+        type=str,
+        default=None,
+        help="Pretrained LoRA path. Required if the training is resumed.",
+    )
+    parser.add_argument(
+        "--use_swanlab",
+        default=False,
+        action="store_true",
+        help="Whether to use SwanLab logger.",
+    )
+    parser.add_argument(
+        "--swanlab_mode",
+        default=None,
+        help="SwanLab mode (cloud or local).",
+    )
+    args = parser.parse_args()
+    return args
+
+
+def data_process(args):
+    dataset = TextVideoDataset(
+        args.dataset_path,
+        os.path.join(args.dataset_path, "metadata.csv") if args.metadata_path is None else args.metadata_path,
+        max_num_frames=args.num_frames,
+        frame_interval=1,
+        num_frames=args.num_frames,
+        height=args.height,
+        width=args.width,
+        is_i2v=args.image_encoder_path is not None,
+        target_fps=args.target_fps,
+    )
+    dataloader = torch.utils.data.DataLoader(
+        dataset,
+        shuffle=False,
+        batch_size=1,
+        num_workers=args.dataloader_num_workers,
+        collate_fn=lambda x: x,
+    )
+    model = LightningModelForDataProcess(
+        text_encoder_path=args.text_encoder_path,
+        image_encoder_path=args.image_encoder_path,
+        vae_path=args.vae_path,
+        tiled=args.tiled,
+        tile_size=(args.tile_size_height, args.tile_size_width),
+        tile_stride=(args.tile_stride_height, args.tile_stride_width),
+        redirected_tensor_path=args.redirected_tensor_path,
+    )
+    trainer = pl.Trainer(
+        accelerator="gpu",
+        devices="auto",
+        default_root_dir=args.output_path,
+    )
+    trainer.test(model, dataloader)
+    
+    
+def train(args):
+    dataset = TensorDataset(
+        args.dataset_path,
+        os.path.join(args.dataset_path, "metadata.csv") if args.metadata_path is None else args.metadata_path,
+        steps_per_epoch=args.steps_per_epoch,
+        redirected_tensor_path=args.redirected_tensor_path,
+    )
+    dataloader = torch.utils.data.DataLoader(
+        dataset,
+        shuffle=True,
+        batch_size=1,
+        num_workers=args.dataloader_num_workers
+    )
+    model = LightningModelForTrain(
+        dit_path=args.dit_path,
+        learning_rate=args.learning_rate,
+        train_architecture=args.train_architecture,
+        lora_rank=args.lora_rank,
+        lora_alpha=args.lora_alpha,
+        lora_target_modules=args.lora_target_modules,
+        init_lora_weights=args.init_lora_weights,
+        use_gradient_checkpointing=args.use_gradient_checkpointing,
+        use_gradient_checkpointing_offload=args.use_gradient_checkpointing_offload,
+        pretrained_lora_path=args.pretrained_lora_path,
+    )
+    if args.use_swanlab:
+        from swanlab.integration.pytorch_lightning import SwanLabLogger
+        swanlab_config = {"UPPERFRAMEWORK": "DiffSynth-Studio"}
+        swanlab_config.update(vars(args))
+        swanlab_logger = SwanLabLogger(
+            project="wan", 
+            name="wan",
+            config=swanlab_config,
+            mode=args.swanlab_mode,
+            logdir=os.path.join(args.output_path, "swanlog"),
+        )
+        logger = [swanlab_logger]
+    else:
+        logger = None
+    trainer = pl.Trainer(
+        max_epochs=args.max_epochs,
+        accelerator="gpu",
+        devices="auto",
+        precision="bf16",
+        strategy=args.training_strategy,
+        default_root_dir=args.output_path,
+        accumulate_grad_batches=args.accumulate_grad_batches,
+        callbacks=[pl.pytorch.callbacks.ModelCheckpoint(save_top_k=-1)],
+        logger=logger,
+    )
+    trainer.fit(model, dataloader)
+
+
+if __name__ == '__main__':
+    args = parse_args()
+    if args.task == "data_process":
+        data_process(args)
+    elif args.task == "train":
+        train(args)

examples/wanvideo/train_wan_t2v_motion.py

@@ -0,0 +1,691 @@
+import torch, os, imageio, argparse
+from torchvision.transforms import v2
+from einops import rearrange
+import lightning as pl
+import pandas as pd
+from diffsynth import WanVideoPipeline, ModelManager, load_state_dict
+from diffsynth.models.wan_video_motion_controller import WanMotionControllerModel
+from diffsynth.pipelines.wan_video import model_fn_wan_video
+from peft import LoraConfig, inject_adapter_in_model
+import torchvision
+from PIL import Image
+import numpy as np
+from tqdm import tqdm
+
+
+
+class TextVideoDataset(torch.utils.data.Dataset):
+    def __init__(self, base_path, metadata_path, max_num_frames=81, frame_interval=1, num_frames=81, height=480, width=832, is_i2v=False, target_fps=None):
+        metadata = pd.read_csv(metadata_path)
+        self.path = [os.path.join(base_path, file_name) for file_name in metadata["file_name"]]
+        self.text = metadata["text"].to_list()
+        
+        self.max_num_frames = max_num_frames
+        self.frame_interval = frame_interval
+        self.num_frames = num_frames
+        self.height = height
+        self.width = width
+        self.is_i2v = is_i2v
+        self.target_fps = target_fps
+            
+        self.frame_process = v2.Compose([
+            v2.CenterCrop(size=(height, width)),
+            v2.Resize(size=(height, width), antialias=True),
+            v2.ToTensor(),
+            v2.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5]),
+        ])
+        
+        
+    def crop_and_resize(self, image):
+        width, height = image.size
+        scale = max(self.width / width, self.height / height)
+        image = torchvision.transforms.functional.resize(
+            image,
+            (round(height*scale), round(width*scale)),
+            interpolation=torchvision.transforms.InterpolationMode.BILINEAR
+        )
+        return image
+
+
+    def load_frames_using_imageio(self, file_path, max_num_frames, start_frame_id, interval, num_frames, frame_process):
+        reader = imageio.get_reader(file_path)
+        if reader.count_frames() < max_num_frames or reader.count_frames() - 1 < start_frame_id + (num_frames - 1) * interval:
+            reader.close()
+            return None
+        
+        frames = []
+        first_frame = None
+        for frame_id in range(num_frames):
+            frame = reader.get_data(start_frame_id + frame_id * interval)
+            frame = Image.fromarray(frame)
+            frame = self.crop_and_resize(frame)
+            if first_frame is None:
+                first_frame = np.array(frame)
+            frame = frame_process(frame)
+            frames.append(frame)
+        reader.close()
+
+        frames = torch.stack(frames, dim=0)
+        frames = rearrange(frames, "T C H W -> C T H W")
+
+        if self.is_i2v:
+            return frames, first_frame
+        else:
+            return frames
+
+
+    def load_video(self, file_path):
+        start_frame_id = 0
+        if self.target_fps is None:
+            frame_interval = self.frame_interval
+        else:
+            reader = imageio.get_reader(file_path)
+            fps = reader.get_meta_data()["fps"]
+            reader.close()
+            frame_interval = max(round(fps / self.target_fps), 1)
+        frames = self.load_frames_using_imageio(file_path, self.max_num_frames, start_frame_id, frame_interval, self.num_frames, self.frame_process)
+        return frames
+    
+    
+    def is_image(self, file_path):
+        file_ext_name = file_path.split(".")[-1]
+        if file_ext_name.lower() in ["jpg", "jpeg", "png", "webp"]:
+            return True
+        return False
+    
+    
+    def load_image(self, file_path):
+        frame = Image.open(file_path).convert("RGB")
+        frame = self.crop_and_resize(frame)
+        first_frame = frame
+        frame = self.frame_process(frame)
+        frame = rearrange(frame, "C H W -> C 1 H W")
+        return frame
+
+
+    def __getitem__(self, data_id):
+        text = self.text[data_id]
+        path = self.path[data_id]
+        try:
+            if self.is_image(path):
+                if self.is_i2v:
+                    raise ValueError(f"{path} is not a video. I2V model doesn't support image-to-image training.")
+                video = self.load_image(path)
+            else:
+                video = self.load_video(path)
+            if self.is_i2v:
+                video, first_frame = video
+                data = {"text": text, "video": video, "path": path, "first_frame": first_frame}
+            else:
+                data = {"text": text, "video": video, "path": path}
+        except:
+            data = None
+        return data
+    
+
+    def __len__(self):
+        return len(self.path)
+
+
+
+class LightningModelForDataProcess(pl.LightningModule):
+    def __init__(self, text_encoder_path, vae_path, image_encoder_path=None, tiled=False, tile_size=(34, 34), tile_stride=(18, 16), redirected_tensor_path=None):
+        super().__init__()
+        model_path = [text_encoder_path, vae_path]
+        if image_encoder_path is not None:
+            model_path.append(image_encoder_path)
+        model_manager = ModelManager(torch_dtype=torch.bfloat16, device="cpu")
+        model_manager.load_models(model_path)
+        self.pipe = WanVideoPipeline.from_model_manager(model_manager)
+
+        self.tiler_kwargs = {"tiled": tiled, "tile_size": tile_size, "tile_stride": tile_stride}
+        self.redirected_tensor_path = redirected_tensor_path
+        
+    def test_step(self, batch, batch_idx):
+        data = batch[0]
+        if data is None or data["video"] is None:
+            return
+        text, video, path = data["text"], data["video"].unsqueeze(0), data["path"]
+        
+        self.pipe.device = self.device
+        if video is not None:
+            # prompt
+            prompt_emb = self.pipe.encode_prompt(text)
+            # video
+            video = video.to(dtype=self.pipe.torch_dtype, device=self.pipe.device)
+            latents = self.pipe.encode_video(video, **self.tiler_kwargs)[0]
+            # image
+            if "first_frame" in batch:
+                first_frame = Image.fromarray(batch["first_frame"][0].cpu().numpy())
+                _, _, num_frames, height, width = video.shape
+                image_emb = self.pipe.encode_image(first_frame, num_frames, height, width)
+            else:
+                image_emb = {}
+            data = {"latents": latents, "prompt_emb": prompt_emb, "image_emb": image_emb}
+            if self.redirected_tensor_path is not None:
+                path = path.replace("/", "_").replace("\\", "_")
+                path = os.path.join(self.redirected_tensor_path, path)
+            torch.save(data, path + ".tensors.pth")
+
+
+
+class TensorDataset(torch.utils.data.Dataset):
+    def __init__(self, base_path, metadata_path=None, steps_per_epoch=1000, redirected_tensor_path=None):
+        if os.path.exists(metadata_path):
+            metadata = pd.read_csv(metadata_path)
+            self.path = [os.path.join(base_path, "train", file_name) for file_name in metadata["file_name"]]
+            print(len(self.path), "videos in metadata.")
+            if redirected_tensor_path is None:
+                self.path = [i + ".tensors.pth" for i in self.path if os.path.exists(i + ".tensors.pth")]
+            else:
+                cached_path = []
+                for path in self.path:
+                    path = path.replace("/", "_").replace("\\", "_")
+                    path = os.path.join(redirected_tensor_path, path)
+                    if os.path.exists(path + ".tensors.pth"):
+                        cached_path.append(path + ".tensors.pth")
+                self.path = cached_path
+        else:
+            print("Cannot find metadata.csv. Trying to search for tensor files.")
+            self.path = [os.path.join(base_path, i) for i in os.listdir(base_path) if i.endswith(".tensors.pth")]
+        print(len(self.path), "tensors cached in metadata.")
+        assert len(self.path) > 0
+        
+        self.steps_per_epoch = steps_per_epoch
+        self.redirected_tensor_path = redirected_tensor_path
+
+
+    def __getitem__(self, index):
+        while True:
+            try:
+                data_id = torch.randint(0, len(self.path), (1,))[0]
+                data_id = (data_id + index) % len(self.path) # For fixed seed.
+                path = self.path[data_id]
+                data = torch.load(path, weights_only=True, map_location="cpu")
+                return data
+            except:
+                continue
+    
+
+    def __len__(self):
+        return self.steps_per_epoch
+
+
+
+class LightningModelForTrain(pl.LightningModule):
+    def __init__(
+        self,
+        dit_path,
+        learning_rate=1e-5,
+        lora_rank=4, lora_alpha=4, train_architecture="lora", lora_target_modules="q,k,v,o,ffn.0,ffn.2", init_lora_weights="kaiming",
+        use_gradient_checkpointing=True, use_gradient_checkpointing_offload=False,
+        pretrained_lora_path=None
+    ):
+        super().__init__()
+        model_manager = ModelManager(torch_dtype=torch.bfloat16, device="cpu")
+        if os.path.isfile(dit_path):
+            model_manager.load_models([dit_path])
+        else:
+            dit_path = dit_path.split(",")
+            model_manager.load_models([dit_path])
+        
+        self.pipe = WanVideoPipeline.from_model_manager(model_manager)
+        self.pipe.scheduler.set_timesteps(1000, training=True)
+        self.freeze_parameters()
+        
+        self.pipe.motion_controller = WanMotionControllerModel().to(torch.bfloat16)
+        self.pipe.motion_controller.init()
+        self.pipe.motion_controller.requires_grad_(True)
+        self.pipe.motion_controller.train()
+        self.motion_bucket_manager = MotionBucketManager()
+        
+        self.learning_rate = learning_rate
+        self.use_gradient_checkpointing = use_gradient_checkpointing
+        self.use_gradient_checkpointing_offload = use_gradient_checkpointing_offload
+        
+        
+    def freeze_parameters(self):
+        # Freeze parameters
+        self.pipe.requires_grad_(False)
+        self.pipe.eval()
+        self.pipe.dit.train()
+        
+        
+    def add_lora_to_model(self, model, lora_rank=4, lora_alpha=4, lora_target_modules="q,k,v,o,ffn.0,ffn.2", init_lora_weights="kaiming", pretrained_lora_path=None, state_dict_converter=None):
+        # Add LoRA to UNet
+        self.lora_alpha = lora_alpha
+        if init_lora_weights == "kaiming":
+            init_lora_weights = True
+            
+        lora_config = LoraConfig(
+            r=lora_rank,
+            lora_alpha=lora_alpha,
+            init_lora_weights=init_lora_weights,
+            target_modules=lora_target_modules.split(","),
+        )
+        model = inject_adapter_in_model(lora_config, model)
+        for param in model.parameters():
+            # Upcast LoRA parameters into fp32
+            if param.requires_grad:
+                param.data = param.to(torch.float32)
+                
+        # Lora pretrained lora weights
+        if pretrained_lora_path is not None:
+            state_dict = load_state_dict(pretrained_lora_path)
+            if state_dict_converter is not None:
+                state_dict = state_dict_converter(state_dict)
+            missing_keys, unexpected_keys = model.load_state_dict(state_dict, strict=False)
+            all_keys = [i for i, _ in model.named_parameters()]
+            num_updated_keys = len(all_keys) - len(missing_keys)
+            num_unexpected_keys = len(unexpected_keys)
+            print(f"{num_updated_keys} parameters are loaded from {pretrained_lora_path}. {num_unexpected_keys} parameters are unexpected.")
+    
+
+    def training_step(self, batch, batch_idx):
+        # Data
+        latents = batch["latents"].to(self.device)
+        prompt_emb = batch["prompt_emb"]
+        prompt_emb["context"] = prompt_emb["context"][0].to(self.device)
+        image_emb = batch["image_emb"]
+        if "clip_feature" in image_emb:
+            image_emb["clip_feature"] = image_emb["clip_feature"][0].to(self.device)
+        if "y" in image_emb:
+            image_emb["y"] = image_emb["y"][0].to(self.device)
+
+        # Loss
+        self.pipe.device = self.device
+        noise = torch.randn_like(latents)
+        timestep_id = torch.randint(0, self.pipe.scheduler.num_train_timesteps, (1,))
+        timestep = self.pipe.scheduler.timesteps[timestep_id].to(dtype=self.pipe.torch_dtype, device=self.pipe.device)
+        extra_input = self.pipe.prepare_extra_input(latents)
+        noisy_latents = self.pipe.scheduler.add_noise(latents, noise, timestep)
+        training_target = self.pipe.scheduler.training_target(latents, noise, timestep)
+        motion_bucket_id = self.motion_bucket_manager(latents)
+        motion_bucket_kwargs = self.pipe.prepare_motion_bucket_id(motion_bucket_id)
+
+        # Compute loss
+        noise_pred = model_fn_wan_video(
+            dit=self.pipe.dit, motion_controller=self.pipe.motion_controller,
+            x=noisy_latents, timestep=timestep, **prompt_emb, **extra_input, **image_emb, **motion_bucket_kwargs,
+            use_gradient_checkpointing=self.use_gradient_checkpointing,
+            use_gradient_checkpointing_offload=self.use_gradient_checkpointing_offload
+        )
+        loss = torch.nn.functional.mse_loss(noise_pred.float(), training_target.float())
+        loss = loss * self.pipe.scheduler.training_weight(timestep)
+
+        # Record log
+        self.log("train_loss", loss, prog_bar=True)
+        return loss
+
+
+    def configure_optimizers(self):
+        trainable_modules = filter(lambda p: p.requires_grad, self.pipe.motion_controller.parameters())
+        optimizer = torch.optim.AdamW(trainable_modules, lr=self.learning_rate)
+        return optimizer
+    
+
+    def on_save_checkpoint(self, checkpoint):
+        checkpoint.clear()
+        trainable_param_names = list(filter(lambda named_param: named_param[1].requires_grad, self.pipe.motion_controller.named_parameters()))
+        trainable_param_names = set([named_param[0] for named_param in trainable_param_names])
+        state_dict = self.pipe.motion_controller.state_dict()
+        lora_state_dict = {}
+        for name, param in state_dict.items():
+            if name in trainable_param_names:
+                lora_state_dict[name] = param
+        checkpoint.update(lora_state_dict)
+        
+        
+
+class MotionBucketManager:
+    def __init__(self):
+        self.thresholds = [
+            0.093750000, 0.094726562, 0.100585938, 0.100585938, 0.108886719, 0.109375000, 0.118652344, 0.127929688, 0.127929688, 0.130859375,
+            0.133789062, 0.137695312, 0.138671875, 0.138671875, 0.139648438, 0.143554688, 0.143554688, 0.147460938, 0.149414062, 0.149414062,
+            0.152343750, 0.153320312, 0.154296875, 0.154296875, 0.157226562, 0.163085938, 0.163085938, 0.164062500, 0.165039062, 0.166992188,
+            0.173828125, 0.179687500, 0.180664062, 0.184570312, 0.187500000, 0.188476562, 0.188476562, 0.189453125, 0.189453125, 0.202148438,
+            0.206054688, 0.210937500, 0.210937500, 0.211914062, 0.214843750, 0.214843750, 0.216796875, 0.216796875, 0.216796875, 0.218750000,
+            0.218750000, 0.221679688, 0.222656250, 0.227539062, 0.229492188, 0.230468750, 0.236328125, 0.243164062, 0.243164062, 0.245117188,
+            0.253906250, 0.253906250, 0.255859375, 0.259765625, 0.275390625, 0.275390625, 0.277343750, 0.279296875, 0.279296875, 0.279296875,
+            0.292968750, 0.292968750, 0.302734375, 0.306640625, 0.312500000, 0.312500000, 0.326171875, 0.330078125, 0.332031250, 0.332031250,
+            0.337890625, 0.343750000, 0.343750000, 0.351562500, 0.355468750, 0.357421875, 0.361328125, 0.367187500, 0.382812500, 0.388671875,
+            0.392578125, 0.392578125, 0.392578125, 0.404296875, 0.404296875, 0.425781250, 0.433593750, 0.507812500, 0.519531250, 0.539062500,
+        ]
+
+    def get_motion_score(self, frames):
+        score = frames[:, :, 1:, :, :].std(dim=2).mean().tolist()
+        return score
+    
+    def get_bucket_id(self, motion_score):
+        for bucket_id in range(len(self.thresholds) - 1):
+            if self.thresholds[bucket_id + 1] > motion_score:
+                return bucket_id
+        return len(self.thresholds)
+
+    def __call__(self, frames):
+        score = self.get_motion_score(frames)
+        bucket_id = self.get_bucket_id(score)
+        return bucket_id
+
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(description="Simple example of a training script.")
+    parser.add_argument(
+        "--task",
+        type=str,
+        default="data_process",
+        required=True,
+        choices=["data_process", "train"],
+        help="Task. `data_process` or `train`.",
+    )
+    parser.add_argument(
+        "--dataset_path",
+        type=str,
+        default=None,
+        required=True,
+        help="The path of the Dataset.",
+    )
+    parser.add_argument(
+        "--output_path",
+        type=str,
+        default="./",
+        help="Path to save the model.",
+    )
+    parser.add_argument(
+        "--metadata_path",
+        type=str,
+        default=None,
+        help="Path to metadata.csv.",
+    )
+    parser.add_argument(
+        "--redirected_tensor_path",
+        type=str,
+        default=None,
+        help="Path to save cached tensors.",
+    )
+    parser.add_argument(
+        "--text_encoder_path",
+        type=str,
+        default=None,
+        help="Path of text encoder.",
+    )
+    parser.add_argument(
+        "--image_encoder_path",
+        type=str,
+        default=None,
+        help="Path of image encoder.",
+    )
+    parser.add_argument(
+        "--vae_path",
+        type=str,
+        default=None,
+        help="Path of VAE.",
+    )
+    parser.add_argument(
+        "--dit_path",
+        type=str,
+        default=None,
+        help="Path of DiT.",
+    )
+    parser.add_argument(
+        "--tiled",
+        default=False,
+        action="store_true",
+        help="Whether enable tile encode in VAE. This option can reduce VRAM required.",
+    )
+    parser.add_argument(
+        "--tile_size_height",
+        type=int,
+        default=34,
+        help="Tile size (height) in VAE.",
+    )
+    parser.add_argument(
+        "--tile_size_width",
+        type=int,
+        default=34,
+        help="Tile size (width) in VAE.",
+    )
+    parser.add_argument(
+        "--tile_stride_height",
+        type=int,
+        default=18,
+        help="Tile stride (height) in VAE.",
+    )
+    parser.add_argument(
+        "--tile_stride_width",
+        type=int,
+        default=16,
+        help="Tile stride (width) in VAE.",
+    )
+    parser.add_argument(
+        "--steps_per_epoch",
+        type=int,
+        default=500,
+        help="Number of steps per epoch.",
+    )
+    parser.add_argument(
+        "--num_frames",
+        type=int,
+        default=81,
+        help="Number of frames.",
+    )
+    parser.add_argument(
+        "--target_fps",
+        type=int,
+        default=None,
+        help="Expected FPS for sampling frames.",
+    )
+    parser.add_argument(
+        "--height",
+        type=int,
+        default=480,
+        help="Image height.",
+    )
+    parser.add_argument(
+        "--width",
+        type=int,
+        default=832,
+        help="Image width.",
+    )
+    parser.add_argument(
+        "--dataloader_num_workers",
+        type=int,
+        default=1,
+        help="Number of subprocesses to use for data loading. 0 means that the data will be loaded in the main process.",
+    )
+    parser.add_argument(
+        "--learning_rate",
+        type=float,
+        default=1e-5,
+        help="Learning rate.",
+    )
+    parser.add_argument(
+        "--accumulate_grad_batches",
+        type=int,
+        default=1,
+        help="The number of batches in gradient accumulation.",
+    )
+    parser.add_argument(
+        "--max_epochs",
+        type=int,
+        default=1,
+        help="Number of epochs.",
+    )
+    parser.add_argument(
+        "--lora_target_modules",
+        type=str,
+        default="q,k,v,o,ffn.0,ffn.2",
+        help="Layers with LoRA modules.",
+    )
+    parser.add_argument(
+        "--init_lora_weights",
+        type=str,
+        default="kaiming",
+        choices=["gaussian", "kaiming"],
+        help="The initializing method of LoRA weight.",
+    )
+    parser.add_argument(
+        "--training_strategy",
+        type=str,
+        default="auto",
+        choices=["auto", "deepspeed_stage_1", "deepspeed_stage_2", "deepspeed_stage_3"],
+        help="Training strategy",
+    )
+    parser.add_argument(
+        "--lora_rank",
+        type=int,
+        default=4,
+        help="The dimension of the LoRA update matrices.",
+    )
+    parser.add_argument(
+        "--lora_alpha",
+        type=float,
+        default=4.0,
+        help="The weight of the LoRA update matrices.",
+    )
+    parser.add_argument(
+        "--use_gradient_checkpointing",
+        default=False,
+        action="store_true",
+        help="Whether to use gradient checkpointing.",
+    )
+    parser.add_argument(
+        "--use_gradient_checkpointing_offload",
+        default=False,
+        action="store_true",
+        help="Whether to use gradient checkpointing offload.",
+    )
+    parser.add_argument(
+        "--train_architecture",
+        type=str,
+        default="lora",
+        choices=["lora", "full"],
+        help="Model structure to train. LoRA training or full training.",
+    )
+    parser.add_argument(
+        "--pretrained_lora_path",
+        type=str,
+        default=None,
+        help="Pretrained LoRA path. Required if the training is resumed.",
+    )
+    parser.add_argument(
+        "--use_swanlab",
+        default=False,
+        action="store_true",
+        help="Whether to use SwanLab logger.",
+    )
+    parser.add_argument(
+        "--swanlab_mode",
+        default=None,
+        help="SwanLab mode (cloud or local).",
+    )
+    args = parser.parse_args()
+    return args
+
+
+def data_process(args):
+    dataset = TextVideoDataset(
+        args.dataset_path,
+        os.path.join(args.dataset_path, "metadata.csv") if args.metadata_path is None else args.metadata_path,
+        max_num_frames=args.num_frames,
+        frame_interval=1,
+        num_frames=args.num_frames,
+        height=args.height,
+        width=args.width,
+        is_i2v=args.image_encoder_path is not None,
+        target_fps=args.target_fps,
+    )
+    dataloader = torch.utils.data.DataLoader(
+        dataset,
+        shuffle=False,
+        batch_size=1,
+        num_workers=args.dataloader_num_workers,
+        collate_fn=lambda x: x,
+    )
+    model = LightningModelForDataProcess(
+        text_encoder_path=args.text_encoder_path,
+        image_encoder_path=args.image_encoder_path,
+        vae_path=args.vae_path,
+        tiled=args.tiled,
+        tile_size=(args.tile_size_height, args.tile_size_width),
+        tile_stride=(args.tile_stride_height, args.tile_stride_width),
+        redirected_tensor_path=args.redirected_tensor_path,
+    )
+    trainer = pl.Trainer(
+        accelerator="gpu",
+        devices="auto",
+        default_root_dir=args.output_path,
+    )
+    trainer.test(model, dataloader)
+    
+    
+def get_motion_thresholds(dataloader):
+    scores = []
+    for data in tqdm(dataloader):
+        scores.append(data["latents"][:, :, 1:, :, :].std(dim=2).mean().tolist())
+    scores = sorted(scores)
+    for i in range(100):
+        s = scores[int(i/100 * len(scores))]
+        print("%.9f" % s, end=", ")
+    
+    
+def train(args):
+    dataset = TensorDataset(
+        args.dataset_path,
+        os.path.join(args.dataset_path, "metadata.csv") if args.metadata_path is None else args.metadata_path,
+        steps_per_epoch=args.steps_per_epoch,
+        redirected_tensor_path=args.redirected_tensor_path,
+    )
+    dataloader = torch.utils.data.DataLoader(
+        dataset,
+        shuffle=True,
+        batch_size=1,
+        num_workers=args.dataloader_num_workers
+    )
+    model = LightningModelForTrain(
+        dit_path=args.dit_path,
+        learning_rate=args.learning_rate,
+        train_architecture=args.train_architecture,
+        lora_rank=args.lora_rank,
+        lora_alpha=args.lora_alpha,
+        lora_target_modules=args.lora_target_modules,
+        init_lora_weights=args.init_lora_weights,
+        use_gradient_checkpointing=args.use_gradient_checkpointing,
+        use_gradient_checkpointing_offload=args.use_gradient_checkpointing_offload,
+        pretrained_lora_path=args.pretrained_lora_path,
+    )
+    if args.use_swanlab:
+        from swanlab.integration.pytorch_lightning import SwanLabLogger
+        swanlab_config = {"UPPERFRAMEWORK": "DiffSynth-Studio"}
+        swanlab_config.update(vars(args))
+        swanlab_logger = SwanLabLogger(
+            project="wan", 
+            name="wan",
+            config=swanlab_config,
+            mode=args.swanlab_mode,
+            logdir=os.path.join(args.output_path, "swanlog"),
+        )
+        logger = [swanlab_logger]
+    else:
+        logger = None
+    trainer = pl.Trainer(
+        max_epochs=args.max_epochs,
+        accelerator="gpu",
+        devices="auto",
+        precision="bf16",
+        strategy=args.training_strategy,
+        default_root_dir=args.output_path,
+        accumulate_grad_batches=args.accumulate_grad_batches,
+        callbacks=[pl.pytorch.callbacks.ModelCheckpoint(save_top_k=-1)],
+        logger=logger,
+    )
+    trainer.fit(model, dataloader)
+
+
+if __name__ == '__main__':
+    args = parse_args()
+    if args.task == "data_process":
+        data_process(args)
+    elif args.task == "train":
+        train(args)