mmagic.models.editors.animatediff.animatediff 源代码

# Copyright (c) OpenMMLab. All rights reserved.
import inspect
from copy import deepcopy
from typing import Dict, List, Optional, Union

import torch
import torch.nn as nn
from einops import rearrange
from mmengine import print_log
from mmengine.logging import MMLogger
from mmengine.model import BaseModel
from safetensors import safe_open
from tqdm import tqdm

from mmagic.models.archs import TokenizerWrapper, set_lora
from mmagic.models.utils import build_module, set_tomesd, set_xformers
from mmagic.registry import DIFFUSION_SCHEDULERS, MODELS
from mmagic.structures import DataSample
from mmagic.utils.typing import SampleList
from .animatediff_utils import (convert_ldm_clip_checkpoint,

[文档]logger = MMLogger.get_current_instance()
[文档]ModelType = Union[Dict, nn.Module]
@MODELS.register_module('animatediff') @MODELS.register_module()
[文档]class AnimateDiff(BaseModel): """Implementation of `AnimateDiff. <>`_ (AnimateDiff). Args: vae (Union[dict, nn.Module]): The config or module for VAE model. text_encoder (Union[dict, nn.Module]): The config or module for text encoder. tokenizer (str): The **name** for CLIP tokenizer. unet (Union[dict, nn.Module]): The config or module for Unet model. schedule (Union[dict, nn.Module]): The config or module for diffusion scheduler. test_scheduler (Union[dict, nn.Module], optional): The config or module for diffusion scheduler in test stage (`self.infer`). If not passed, will use the same scheduler as `schedule`. Defaults to None. lora_config (dict, optional): The config for LoRA finetuning. Defaults to None. val_prompts (Union[str, List[str]], optional): The prompts for validation. Defaults to None. class_prior_prompt (str, optional): The prompt for class prior loss. num_class_images (int, optional): The number of images for class prior. Defaults to 3. prior_loss_weight (float, optional): The weight for class prior loss. Defaults to 0. fine_tune_text_encoder (bool, optional): Whether to fine-tune text encoder. Defaults to False. dtype (str, optional): The dtype for the model. Defaults to 'fp16'. enable_xformers (bool, optional): Whether to use xformers. Defaults to True. noise_offset_weight (bool, optional): The weight of noise offset introduced in # noqa Defaults to 0. tomesd_cfg (dict, optional): The config for TOMESD. Please refers to and for detail. # noqa Defaults to None. data_preprocessor (dict, optional): The pre-process config of :class:`BaseDataPreprocessor`. Defaults to dict(type='DataPreprocessor'). init_cfg (dict, optional): The weight initialized config for :class:`BaseModule`. Defaults to None/ """ def __init__( self, vae: ModelType, text_encoder: ModelType, tokenizer: str, unet: ModelType, scheduler: ModelType, test_scheduler: Optional[ModelType] = None, dtype: str = 'fp32', enable_xformers: bool = True, noise_offset_weight: float = 0, tomesd_cfg: Optional[dict] = None, data_preprocessor=dict(type='DataPreprocessor'), motion_module_cfg: Optional[dict] = None, dream_booth_lora_cfg: Optional[dict] = None, ): super().__init__(data_preprocessor) default_args = dict() if dtype is not None: default_args['dtype'] = dtype self.dtype = torch.float32 if dtype in ['float16', 'fp16', 'half']: self.dtype = torch.float16 elif dtype == 'bf16': self.dtype = torch.bfloat16 else: assert dtype in [ 'fp32', None ], ('dtype must be one of \'fp32\', \'fp16\', \'bf16\' or None.') self.vae = build_module(vae, MODELS, default_args=default_args) self.unet = build_module(unet, MODELS) # NOTE: initialize unet as fp32 self._unet_ori_dtype = next(self.unet.parameters()).dtype print_log(f'Set UNet dtype to \'{self._unet_ori_dtype}\'.', 'current') self.init_motion_module(motion_module_cfg) self.scheduler = build_module(scheduler, DIFFUSION_SCHEDULERS) if test_scheduler is None: self.test_scheduler = deepcopy(self.scheduler) else: self.test_scheduler = build_module(test_scheduler, DIFFUSION_SCHEDULERS) self.text_encoder = build_module(text_encoder, MODELS) if not isinstance(tokenizer, str): self.tokenizer = tokenizer else: # NOTE: here we assume tokenizer is an string self.tokenizer = TokenizerWrapper(tokenizer, subfolder='tokenizer') self.unet_sample_size = self.unet.sample_size self.vae_scale_factor = 2**(len(self.vae.block_out_channels) - 1) self.enable_noise_offset = noise_offset_weight > 0 self.noise_offset_weight = noise_offset_weight self.enable_xformers = enable_xformers self.unet.set_use_memory_efficient_attention_xformers(True) self.tomesd_cfg = tomesd_cfg self.set_tomesd() self.init_dreambooth_lora(dream_booth_lora_cfg) self.prepare_model()
[文档] def set_xformers(self, module: Optional[nn.Module] = None) -> nn.Module: """Set xformers for the model. Returns: nn.Module: The model with xformers. """ if self.enable_xformers: if module is None: set_xformers(self) else: set_xformers(module)
[文档] def set_tomesd(self) -> nn.Module: """Set ToMe for the stable diffusion model. Returns: nn.Module: The model with ToMe. """ if self.tomesd_cfg is not None: set_tomesd(self, **self.tomesd_cfg)
[文档] def device(self): """Set device for the model.""" return next(self.parameters()).device
[文档] def init_motion_module(self, motion_module_cfg): if motion_module_cfg is not None: if 'path' in motion_module_cfg.keys(): motion_module_state_dict = torch.load( motion_module_cfg['path'], map_location='cpu') # if "global_step" in motion_module_state_dict: # func_args.update({"global_step": # motion_module_state_dict["global_step"]}) missing, unexpected = self.unet.load_state_dict( motion_module_state_dict, strict=False) assert len(unexpected) == 0
[文档] def init_dreambooth_lora(self, dream_booth_lora_cfg): # TODO: finish if dream_booth_lora_cfg is not None: if 'path' in dream_booth_lora_cfg.keys(): state_dict = {} with safe_open( dream_booth_lora_cfg['path'], framework='pt', device='cpu') as f: for key in f.keys(): state_dict[key] = f.get_tensor(key) # vae converted_vae_checkpoint = convert_ldm_vae_checkpoint( state_dict, self.vae.config) self.vae.load_state_dict(converted_vae_checkpoint) # unet converted_unet_checkpoint = convert_ldm_unet_checkpoint( state_dict, self.unet.config) self.unet.load_state_dict( converted_unet_checkpoint, strict=False) # text_model self.text_encoder = convert_ldm_clip_checkpoint(state_dict)
# self.convert_lora(state_dict)
[文档] def _encode_prompt(self, prompt, device, num_videos_per_prompt, do_classifier_free_guidance, negative_prompt): """Encodes the prompt into text encoder hidden states.""" batch_size = len(prompt) if isinstance(prompt, list) else 1 text_inputs = self.tokenizer( prompt, padding='max_length', max_length=self.tokenizer.model_max_length, truncation=True, return_tensors='pt', ) text_input_ids = text_inputs.input_ids untruncated_ids = self.tokenizer( prompt, padding='longest', return_tensors='pt').input_ids if untruncated_ids.shape[-1] >= text_input_ids.shape[ -1] and not torch.equal(text_input_ids, untruncated_ids): removed_text = self.tokenizer.batch_decode( untruncated_ids[:, self.tokenizer.model_max_length - 1:-1]) logger.warning( 'The following part of your input was truncated ' f'because CLIP can only handle sequences up to' f' {self.tokenizer.model_max_length} tokens: {removed_text}') text_encoder = self.text_encoder.module if hasattr( self.text_encoder, 'module') else self.text_encoder if hasattr(text_encoder.config, 'use_attention_mask' ) and text_encoder.config.use_attention_mask: attention_mask = else: attention_mask = None text_embeddings = self.text_encoder(, attention_mask=attention_mask, ) text_embeddings = text_embeddings[0] # duplicate text embeddings for each generation # per prompt, using mps friendly method bs_embed, seq_len, _ = text_embeddings.shape text_embeddings = text_embeddings.repeat(1, num_videos_per_prompt, 1) text_embeddings = text_embeddings.view( bs_embed * num_videos_per_prompt, seq_len, -1) # get unconditional embeddings for classifier free guidance if do_classifier_free_guidance: uncond_tokens: List[str] if negative_prompt is None: uncond_tokens = [''] * batch_size elif type(prompt) is not type(negative_prompt): raise TypeError( f'`negative_prompt` should be the same type ' f'to `prompt`, but got {type(negative_prompt)} !=' f' {type(prompt)}.') elif isinstance(negative_prompt, str): uncond_tokens = [negative_prompt] elif batch_size != len(negative_prompt): raise ValueError( f'`negative_prompt`: {negative_prompt} has ' f'batch size {len(negative_prompt)}, but `prompt`:' f' {prompt} has batch size {batch_size}. Please ' f'make sure that passed `negative_prompt` matches' ' the batch size of `prompt`.') else: uncond_tokens = negative_prompt max_length = text_input_ids.shape[-1] uncond_input = self.tokenizer( uncond_tokens, padding='max_length', max_length=max_length, truncation=True, return_tensors='pt', ) if hasattr(text_encoder.config, 'use_attention_mask' ) and text_encoder.config.use_attention_mask: attention_mask = else: attention_mask = None uncond_embeddings = self.text_encoder(, attention_mask=attention_mask, ) uncond_embeddings = uncond_embeddings[0] # duplicate unconditional embeddings for each generation # per prompt, using mps friendly method seq_len = uncond_embeddings.shape[1] uncond_embeddings = uncond_embeddings.repeat( 1, num_videos_per_prompt, 1) uncond_embeddings = uncond_embeddings.view( batch_size * num_videos_per_prompt, seq_len, -1) # For classifier free guidance, we need to do two forward passes. # Here we concatenate the unconditional and text embeddings # into a single batch to avoid doing two forward passes text_embeddings =[uncond_embeddings, text_embeddings]) return text_embeddings
[文档] def decode_latents(self, latents): """latents decoder.""" video_length = latents.shape[2] latents = 1 / 0.18215 * latents latents = rearrange(latents, 'b c f h w -> (b f) c h w') # video = self.vae.decode(latents).sample video = [] for frame_idx in tqdm(range(latents.shape[0])): video.append( self.vae.decode(latents[frame_idx:frame_idx + 1]).sample) video = video = rearrange(video, '(b f) c h w -> b c f h w', f=video_length) video = (video / 2 + 0.5).clamp(0, 1) # we always cast to float32 as this does not cause significant # overhead and is compatible with bfloa16 video = video.cpu().float().numpy() return video
[文档] def prepare_extra_step_kwargs(self, generator, eta): """Prepare extra kwargs for the scheduler step, since not all schedulers have the same signature eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers.""" # prepare extra kwargs for the scheduler step, since not all # schedulers have the same signature # eta (η) is only used with the DDIMScheduler, it will # be ignored for other schedulers. # eta corresponds to η in DDIM paper: # and should be between [0, 1] accepts_eta = 'eta' in set( inspect.signature(self.scheduler.step).parameters.keys()) extra_step_kwargs = {} if accepts_eta: extra_step_kwargs['eta'] = eta # check if the scheduler accepts generator accepts_generator = 'generator' in set( inspect.signature(self.scheduler.step).parameters.keys()) if accepts_generator: extra_step_kwargs['generator'] = generator return extra_step_kwargs
[文档] def check_inputs(self, prompt, height, width): """Check inputs. Raise error if not correct """ if not isinstance(prompt, str) and not isinstance(prompt, list): raise ValueError(f'`prompt` has to be of type `str`' f' or `list` but is {type(prompt)}') if height % 8 != 0 or width % 8 != 0: raise ValueError(f'`height` and `width` have to be divisible'
f' by 8 but are {height} and {width}.') # if (callback_steps is None) or ( # callback_steps is not None and (not isinstance(callback_steps, # int) or callback_steps <= 0) # ): # raise ValueError( # f"`callback_steps` has to be a positive integer but # is {callback_steps} of type" # f" {type(callback_steps)}." # )
[文档] def convert_lora(self, state_dict, LORA_PREFIX_UNET='lora_unet', LORA_PREFIX_TEXT_ENCODER='lora_te', alpha=0.6): """ Convert lora for unet and text_encoder TODO: use this function to convert lora Args: state_dict (_type_): _description_ LORA_PREFIX_UNET (str, optional): _description_. Defaults to 'lora_unet'. LORA_PREFIX_TEXT_ENCODER (str, optional): _description_. Defaults to 'lora_te'. alpha (float, optional): _description_. Defaults to 0.6. Returns: TODO: check each output type _type_: unet && text_encoder """ # load base model # pipeline = StableDiffusionPipeline.from_pretrained(base_model_path, # torch_dtype=torch.float32) # load LoRA weight from .safetensors # state_dict = load_file(checkpoint_path) visited = [] # directly update weight in diffusers model for key in state_dict: # it is suggested to print out the key, it usually # will be something like below # "lora_te_text_model_encoder_layers_0_self_attn_k_proj.lora_down.weight" # as we have set the alpha beforehand, so just skip if '.alpha' in key or key in visited: continue if 'text' in key: layer_infos = key.split('.')[0].split( LORA_PREFIX_TEXT_ENCODER + '_')[-1].split('_') curr_layer = self.text_encoder else: layer_infos = key.split('.')[0].split(LORA_PREFIX_UNET + '_')[-1].split('_') curr_layer = self.unet # find the target layer temp_name = layer_infos.pop(0) while len(layer_infos) > -1: try: curr_layer = curr_layer.__getattr__(temp_name) if len(layer_infos) > 0: temp_name = layer_infos.pop(0) elif len(layer_infos) == 0: break except Exception: if len(temp_name) > 0: temp_name += '_' + layer_infos.pop(0) else: temp_name = layer_infos.pop(0) pair_keys = [] if 'lora_down' in key: pair_keys.append(key.replace('lora_down', 'lora_up')) pair_keys.append(key) else: pair_keys.append(key) pair_keys.append(key.replace('lora_up', 'lora_down')) # update weight if len(state_dict[pair_keys[0]].shape) == 4: weight_up = state_dict[pair_keys[0]].squeeze(3).squeeze(2).to( torch.float32) weight_down = state_dict[pair_keys[1]].squeeze(3).squeeze( 2).to(torch.float32) += alpha * weight_up, weight_down).unsqueeze(2).unsqueeze(3).to( else: weight_up = state_dict[pair_keys[0]].to(torch.float32) weight_down = state_dict[pair_keys[1]].to(torch.float32) += alpha * weight_up, weight_down).to( # update visited list for item in pair_keys: visited.append(item) return self.unet, self.text_encoder
[文档] def prepare_latents(self, batch_size, num_channels_latents, video_length, height, width, dtype, device, generator, latents=None): """Prepare latent variables.""" shape = (batch_size, num_channels_latents, video_length, height // self.vae_scale_factor, width // self.vae_scale_factor) if isinstance(generator, list) and len(generator) != batch_size: raise ValueError( f'You have passed a list of generators of length ' f'{len(generator)}, but requested an effective batch' f' size of {batch_size}. Make sure the batch size matches the' f' length of the generators.') if latents is None: rand_device = 'cpu' if device.type == 'mps' else device if isinstance(generator, list): shape = shape # shape = (1,) + shape[1:] latents = [ torch.randn( shape, generator=generator[i], device=rand_device, dtype=dtype) for i in range(batch_size) ] latents =, dim=0).to(device) else: latents = torch.randn( shape, generator=generator, device=rand_device, dtype=dtype).to(device) else: if latents.shape != shape: raise ValueError(f'Unexpected latents shape, got ' f'{latents.shape}, expected {shape}') latents = # scale the initial noise by the standard deviation # required by the scheduler latents = latents * self.scheduler.init_noise_sigma return latents
[文档] def prepare_model(self): """Prepare model for training. Move model to target dtype and disable gradient for some models. """ self.vae.requires_grad_(False) print_log('Set VAE untrainable.', 'current') print_log(f'Move VAE to {self.dtype}.', 'current') # if not self.finetune_text_encoder or self.lora_config: if 1: self.text_encoder.requires_grad_(False) print_log('Set Text Encoder untrainable.', 'current') print_log(f'Move Text Encoder to {self.dtype}.', 'current') # if self.lora_config: if 1: self.unet.requires_grad_(False) print_log('Set Unet untrainable.', 'current')
[文档] def set_lora(self): """Set LORA for model.""" if self.lora_config: set_lora(self.unet, self.lora_config)
[文档] def val_step(self, data: dict) -> SampleList: """Gets the generated image of given data. Calls ``self.data_preprocessor`` and ``self.infer`` in order. Return the generated results which will be passed to evaluator or visualizer. Args: data (dict or tuple or list): Data sampled from dataset. Returns: SampleList: Generated image or image dict. """ data = self.data_preprocessor(data) data_samples = data['data_samples'] if self.val_prompts is None: prompt = data_samples.prompt else: prompt = self.val_prompts # construct a fake data_sample for destruct data_samples.split() * len(prompt) data_samples = DataSample.stack(data_samples.split() * len(prompt)) output = self.infer(prompt, return_type='tensor') samples = output['samples'] samples = self.data_preprocessor.destruct(samples, data_samples) out_data_sample = DataSample(fake_img=samples, prompt=prompt) data_sample_list = out_data_sample.split() return data_sample_list
[文档] def test_step(self, data: dict) -> SampleList: """Gets the generated image of given data. Calls ``self.data_preprocessor`` and ``self.infer`` in order. Return the generated results which will be passed to evaluator or visualizer. Args: data (dict or tuple or list): Data sampled from dataset. Returns: SampleList: Generated image or image dict. """ if self.val_prompts is None: data = self.data_preprocessor(data) data_samples = data['data_samples'] prompt = data_samples.prompt else: prompt = self.val_prompts # construct a fake data_sample for destruct data_samples = DataSample.stack(data['data_samples'] * len(prompt)) output = self.infer(prompt, return_type='tensor') samples = output['samples'] samples = self.data_preprocessor.destruct(samples, data_samples) out_data_sample = DataSample(fake_img=samples, prompt=prompt) data_sample_list = out_data_sample.split() return data_sample_list
[文档] def infer(self, prompt: Union[str, List[str]], video_length: Optional[int] = 16, height: Optional[int] = None, width: Optional[int] = None, num_inference_steps: int = 50, guidance_scale: float = 7.5, negative_prompt: Optional[Union[str, List[str]]] = None, num_videos_per_prompt: Optional[int] = 1, eta: float = 0.0, generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None, latents: Optional[torch.FloatTensor] = None, return_type: Optional[str] = 'tensor', show_progress: bool = True, seed: Optional[int] = 1007): """Function invoked when calling the pipeline for generation. Args: prompt (str or List[str]): The prompt or prompts to guide the video generation. video_length (int, Option): The number of frames of the generated video. Defaults to 16. height (int, Optional): The height in pixels of the generated image. If not passed, the height will be `self.unet_sample_size * self.vae_scale_factor` Defaults to None. width (int, Optional): The width in pixels of the generated image. If not passed, the width will be `self.unet_sample_size * self.vae_scale_factor` Defaults to None. num_inference_steps (int): The number of denoising steps. More denoising steps usually lead to a higher quality video at the expense of slower inference. Defaults to 50. guidance_scale (float): Guidance scale as defined in Classifier- Free Diffusion Guidance ( Defaults to 7.5 negative_prompt (str or List[str], optional): The prompt or prompts not to guide the video generation. Ignored when not using guidance (i.e., ignored if `guidance_scale` is less than 1). Defaults to None. num_videos_per_prompt (int): The number of videos to generate per prompt. Defaults to 1. eta (float): Corresponds to parameter eta (η) in the DDIM paper: Only applies to DDIMScheduler, will be ignored for others. Defaults to 0.0. generator (torch.Generator, optional): A torch generator to make generation deterministic. Defaults to None. latents (torch.FloatTensor, optional): Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for video generation. Can be used to tweak the same generation with different prompts. If not provided, a latents tensor will be generated by sampling using the supplied random `generator`. Defaults to None. return_type (str): The return type of the inference results. Supported types are 'video', 'numpy', 'tensor'. If 'video' is passed, a list of PIL images will be returned. If 'numpy' is passed, a numpy array with shape [N, C, H, W] will be returned, and the value range will be same as decoder's output range. If 'tensor' is passed, the decoder's output will be returned. Defaults to 'image'. #TODO Returns: dict: A dict containing the generated video """ assert return_type in ['image', 'tensor', 'numpy'] # 0. Default height and width to unet height = height or self.unet_sample_size * self.vae_scale_factor width = width or self.unet_sample_size * self.vae_scale_factor if seed != -1: torch.manual_seed(seed) print_log(f'current seed: {torch.initial_seed()}') print_log(f'sampling {prompt} ...') # 1. Check inputs. Raise error if not correct self.check_inputs(prompt, height, width) # NOTE: aligned with origin repo # 2. Define call parameters batch_size = 1 if latents is not None: batch_size = latents.shape[0] if isinstance(prompt, list): batch_size = len(prompt) device = self.device # here `guidance_scale` is defined analog to the # guidance weight `w` of equation (2) # of the Imagen paper: . # `guidance_scale = 1` # corresponds to doing no classifier free guidance. do_classifier_free_guidance = guidance_scale > 1.0 video_dtype = self.vae.module.dtype if hasattr(self.vae, 'module') \ else self.vae.dtype # 3. Encode input prompt prompt = prompt if isinstance(prompt, list) else [prompt] * batch_size if negative_prompt is not None: negative_prompt = negative_prompt if isinstance( negative_prompt, list) else [negative_prompt] * batch_size text_embeddings = self._encode_prompt( prompt, device, num_videos_per_prompt, do_classifier_free_guidance, negative_prompt) # NOTE aligned with origin repo # 4. Prepare timesteps # self.scheduler.set_timesteps(num_inference_steps, device=device) self.test_scheduler.set_timesteps(num_inference_steps) timesteps = self.test_scheduler.timesteps # 5. Prepare latent variables num_channels_latents = self.unet.config.in_channels latents = self.prepare_latents( batch_size * num_videos_per_prompt, num_channels_latents, video_length, height, width, text_embeddings.dtype, device, generator, latents, ) # NOTE aligned with origin repo latents_dtype = latents.dtype # 6. Prepare extra step kwargs. extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta) # 7. Denoising loop if show_progress: timesteps = tqdm(timesteps) for i, t in enumerate(timesteps): # expand the latents if we are doing classifier free guidance latent_model_input = [latents] * 2) if do_classifier_free_guidance else latents latent_model_input = self.test_scheduler.scale_model_input( latent_model_input, t) # predict the noise residual noise_pred = self.unet( latent_model_input, t, encoder_hidden_states=text_embeddings, )['sample'].to(dtype=latents_dtype) # perform guidance if do_classifier_free_guidance: noise_pred_uncond, noise_pred_text = noise_pred.chunk(2) noise_pred = noise_pred_uncond + guidance_scale * ( noise_pred_text - noise_pred_uncond) # compute the previous noisy sample x_t -> x_t-1 latents = self.test_scheduler.step( noise_pred, t, latents, **extra_step_kwargs)['prev_sample'] # FIXME: not aligned # FIXME: revise config thresholding=False # scheduler pred_original_sample not aligned # fixed clip_sample=False # 8. Post-processing video = self.decode_latents( if return_type == 'tensor': video = torch.from_numpy(video) return {'samples': video}
[文档] def forward(self, inputs: torch.Tensor, data_samples: Optional[list] = None, mode: str = 'tensor') -> Union[Dict[str, torch.Tensor], list]: """forward is not implemented now.""" raise NotImplementedError( 'Forward is not implemented now, please use infer.')
Read the Docs v: latest
On Read the Docs
Project Home

Free document hosting provided by Read the Docs.