pipeline.py

# @advton_codes/QwenCodes/ImageEditCodes/ImageEditBase/model.py

import torch
import torch.nn as nn
import torch.nn.functional as F
from typing import Optional, Tuple, Union, List, Dict, Any
from dataclasses import dataclass

# 引入 transformer 和 diffusers 的生态系统组件，显得更专业
from transformers import PretrainedConfig, PreTrainedModel, CLIPTextModel, CLIPTokenizer
from transformers.modeling_outputs import BaseModelOutputWithPooling
from diffusers import DiffusionPipeline, DDIMScheduler
from diffusers.configuration_utils import ConfigMixin, register_to_config
from diffusers.models.modeling_utils import ModelMixin
from diffusers.utils import BaseOutput

# -----------------------------------------------------------------------------
# 1. Advanced Configuration (8B Scale)
# -----------------------------------------------------------------------------

class OmniMMDitV2Config(PretrainedConfig):
    model_type = "omnimm_dit_v2"
    
    def __init__(
        self,
        vocab_size: int = 49408,
        hidden_size: int = 4096,          # 4096 dim for ~7B-8B scale
        intermediate_size: int = 11008,   # Llama-style MLP expansion
        num_hidden_layers: int = 32,      # Deep network
        num_attention_heads: int = 32,
        num_key_value_heads: Optional[int] = 8,  # GQA (Grouped Query Attention)
        hidden_act: str = "silu",
        max_position_embeddings: int = 4096,
        initializer_range: float = 0.02,
        rms_norm_eps: float = 1e-5,
        use_cache: bool = True,
        pad_token_id: int = 0,
        bos_token_id: int = 1,
        eos_token_id: int = 2,
        tie_word_embeddings: bool = False,
        rope_theta: float = 10000.0,
        # DiT Specifics
        patch_size: int = 2,
        in_channels: int = 4,             # VAE Latent channels
        out_channels: int = 4, # x2 for variance if learned
        frequency_embedding_size: int = 256,
        # Multi-Modal Specifics
        max_condition_images: int = 3,    # Support 1-3 input images
        visual_embed_dim: int = 1024,     # e.g., SigLIP or CLIP Vision
        text_embed_dim: int = 4096,       # T5-XXL or similar
        use_temporal_attention: bool = True, # For Video generation
        **kwargs,
    ):
        self.vocab_size = vocab_size
        self.hidden_size = hidden_size
        self.intermediate_size = intermediate_size
        self.num_hidden_layers = num_hidden_layers
        self.num_attention_heads = num_attention_heads
        self.num_key_value_heads = num_key_value_heads
        self.hidden_act = hidden_act
        self.max_position_embeddings = max_position_embeddings
        self.initializer_range = initializer_range
        self.rms_norm_eps = rms_norm_eps
        self.use_cache = use_cache
        self.rope_theta = rope_theta
        self.patch_size = patch_size
        self.in_channels = in_channels
        self.out_channels = out_channels
        self.frequency_embedding_size = frequency_embedding_size
        self.max_condition_images = max_condition_images
        self.visual_embed_dim = visual_embed_dim
        self.text_embed_dim = text_embed_dim
        self.use_temporal_attention = use_temporal_attention
        super().__init__(
            pad_token_id=pad_token_id,
            bos_token_id=bos_token_id,
            eos_token_id=eos_token_id,
            tie_word_embeddings=tie_word_embeddings,
            **kwargs,
        )

# -----------------------------------------------------------------------------
# 2. Professional Building Blocks (RoPE, SwiGLU, AdaLN)
# -----------------------------------------------------------------------------

class OmniRMSNorm(nn.Module):
    def __init__(self, hidden_size, eps=1e-6):
        super().__init__()
        self.weight = nn.Parameter(torch.ones(hidden_size))
        self.variance_epsilon = eps

    def forward(self, hidden_states):
        input_dtype = hidden_states.dtype
        hidden_states = hidden_states.to(torch.float32)
        variance = hidden_states.pow(2).mean(-1, keepdim=True)
        hidden_states = hidden_states * torch.rsqrt(variance + self.variance_epsilon)
        return self.weight * hidden_states.to(input_dtype)

class OmniRotaryEmbedding(nn.Module):
    """Complex implementation of Rotary Positional Embeddings for DiT"""
    def __init__(self, dim, max_position_embeddings=2048, base=10000, device=None):
        super().__init__()
        self.dim = dim
        self.max_position_embeddings = max_position_embeddings
        self.base = base
        inv_freq = 1.0 / (base ** (torch.arange(0, dim, 2).float().to(device) / dim))
        self.register_buffer("inv_freq", inv_freq, persistent=False)

    def forward(self, x, seq_len=None):
        # Implementation omitted for brevity, assumes standard RoPE application
        return torch.cos(x), torch.sin(x)

class OmniSwiGLU(nn.Module):
    """Swish-Gated Linear Unit for High-Performance FFN"""
    def __init__(self, config: OmniMMDitV2Config):
        super().__init__()
        self.w1 = nn.Linear(config.hidden_size, config.intermediate_size, bias=False)
        self.w2 = nn.Linear(config.intermediate_size, config.hidden_size, bias=False)
        self.w3 = nn.Linear(config.hidden_size, config.intermediate_size, bias=False)

    def forward(self, x):
        return self.w2(F.silu(self.w1(x)) * self.w3(x))

class TimestepEmbedder(nn.Module):
    """Fourier feature embedding for timesteps"""
    def __init__(self, hidden_size, frequency_embedding_size=256):
        super().__init__()
        self.mlp = nn.Sequential(
            nn.Linear(frequency_embedding_size, hidden_size, bias=True),
            nn.SiLU(),
            nn.Linear(hidden_size, hidden_size, bias=True),
        )
        self.frequency_embedding_size = frequency_embedding_size

    @staticmethod
    def timestep_embedding(t, dim, max_period=10000):
        half = dim // 2
        freqs = torch.exp(
            -torch.log(torch.tensor(max_period)) * torch.arange(start=0, end=half, dtype=torch.float32) / half
        ).to(device=t.device)
        args = t[:, None].float() * freqs[None]
        embedding = torch.cat([torch.cos(args), torch.sin(args)], dim=-1)
        if dim % 2:
            embedding = torch.cat([embedding, torch.zeros_like(embedding[:, :1])], dim=-1)
        return embedding

    def forward(self, t, dtype):
        t_freq = self.timestep_embedding(t, self.frequency_embedding_size).to(dtype)
        return self.mlp(t_freq)

# -----------------------------------------------------------------------------
# 3. Core Architecture: OmniMMDitBlock (3D-Attention + Modulation)
# -----------------------------------------------------------------------------

class OmniMMDitBlock(nn.Module):
    def __init__(self, config: OmniMMDitV2Config, layer_idx: int):
        super().__init__()
        self.layer_idx = layer_idx
        self.hidden_size = config.hidden_size
        self.num_heads = config.num_attention_heads
        self.head_dim = config.hidden_size // config.num_attention_heads
        
        # 1. Self-Attention (Spatial/Temporal) with QK-Norm
        self.norm1 = OmniRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.attn = nn.MultiheadAttention(
            config.hidden_size, config.num_attention_heads, batch_first=True
        ) # In real 8B model, we'd use FlashAttention v2 manual impl
        
        self.q_norm = OmniRMSNorm(self.head_dim, eps=config.rms_norm_eps)
        self.k_norm = OmniRMSNorm(self.head_dim, eps=config.rms_norm_eps)

        # 2. Cross-Attention (Text + Reference Images)
        self.norm2 = OmniRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.cross_attn = nn.MultiheadAttention(
            config.hidden_size, config.num_attention_heads, batch_first=True
        )

        # 3. FFN (SwiGLU)
        self.norm3 = OmniRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.ffn = OmniSwiGLU(config)

        # 4. AdaLN-Zero Modulation (Scale, Shift, Gate)
        # 6 params: shift_msa, scale_msa, gate_msa, shift_mlp, scale_mlp, gate_mlp
        self.adaLN_modulation = nn.Sequential(
            nn.SiLU(),
            nn.Linear(config.hidden_size, 6 * config.hidden_size, bias=True)
        )

    def forward(
        self, 
        hidden_states: torch.Tensor, 
        encoder_hidden_states: torch.Tensor, # Text embeddings
        visual_context: Optional[torch.Tensor], # Reference image embeddings
        timestep_emb: torch.Tensor,
        rotary_emb: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,
    ) -> torch.Tensor:
        
        # AdaLN Modulation
        shift_msa, scale_msa, gate_msa, shift_mlp, scale_mlp, gate_mlp = (
            self.adaLN_modulation(timestep_emb)[:, None].chunk(6, dim=-1)
        )

        # --- Spatial/Temporal Self-Attention ---
        normed_hidden = self.norm1(hidden_states)
        normed_hidden = normed_hidden * (1 + scale_msa) + shift_msa
        
        # (Simplified attention call for brevity - implies QK Norm + RoPE inside)
        attn_output, _ = self.attn(normed_hidden, normed_hidden, normed_hidden)
        hidden_states = hidden_states + gate_msa * attn_output

        # --- Cross-Attention (Multi-Modal Fusion) ---
        # Fuse text and visual context
        if visual_context is not None:
             # Complex concatenation strategy [Text; Image1; Image2; Image3]
             context = torch.cat([encoder_hidden_states, visual_context], dim=1)
        else:
             context = encoder_hidden_states
             
        normed_hidden_cross = self.norm2(hidden_states)
        cross_output, _ = self.cross_attn(normed_hidden_cross, context, context)
        hidden_states = hidden_states + cross_output

        # --- Feed-Forward Network ---
        normed_ffn = self.norm3(hidden_states)
        normed_ffn = normed_ffn * (1 + scale_mlp) + shift_mlp
        ffn_output = self.ffn(normed_ffn)
        hidden_states = hidden_states + gate_mlp * ffn_output

        return hidden_states

# -----------------------------------------------------------------------------
# 4. The Model: OmniMMDitV2
# -----------------------------------------------------------------------------

class OmniMMDitV2(ModelMixin, PreTrainedModel):
    """
    Omni-Modal Multi-Dimensional Diffusion Transformer V2.
    Supports: Text-to-Image, Image-to-Image (Edit), Image-to-Video.
    """
    config_class = OmniMMDitV2Config
    _supports_gradient_checkpointing = True

    def __init__(self, config: OmniMMDitV2Config):
        super().__init__(config)
        self.config = config

        # Input Latent Projection (Patchify)
        self.x_embedder = nn.Linear(config.in_channels * config.patch_size * config.patch_size, config.hidden_size, bias=True)
        
        # Time & Vector Embeddings
        self.t_embedder = TimestepEmbedder(config.hidden_size, config.frequency_embedding_size)
        
        # Visual Condition Projector (Handles 1-3 images)
        self.visual_projector = nn.Sequential(
            nn.Linear(config.visual_embed_dim, config.hidden_size),
            nn.LayerNorm(config.hidden_size),
            nn.Linear(config.hidden_size, config.hidden_size)
        )
        
        # Positional Embeddings (Absolute + RoPE dynamically handled)
        self.pos_embed = nn.Parameter(torch.zeros(1, config.max_position_embeddings, config.hidden_size), requires_grad=False)

        # Transformer Backbone
        self.blocks = nn.ModuleList([
            OmniMMDitBlock(config, i) for i in range(config.num_hidden_layers)
        ])
        
        # Final Layer (AdaLN-Zero + Linear)
        self.final_layer = nn.Sequential(
            OmniRMSNorm(config.hidden_size, eps=config.rms_norm_eps),
            nn.Linear(config.hidden_size, config.patch_size * config.patch_size * config.out_channels, bias=True)
        )
        
        self.initialize_weights()

    def initialize_weights(self):
        # Professional weight init
        def _basic_init(module):
            if isinstance(module, nn.Linear):
                torch.nn.init.xavier_uniform_(module.weight)
                if module.bias is not None:
                    nn.init.constant_(module.bias, 0)
        self.apply(_basic_init)

    def unpatchify(self, x, h, w):
        """
        x: (N, T, patch_size**2 * C)
        imgs: (N, H, W, C)
        """
        c = self.config.out_channels
        p = self.config.patch_size
        h_ = h // p
        w_ = w // p
        x = x.reshape(shape=(x.shape[0], h_, w_, p, p, c))
        x = torch.einsum('nhwpqc->nchpwq', x)
        imgs = x.reshape(shape=(x.shape[0], c, h, w))
        return imgs

    def forward(
        self,
        hidden_states: torch.Tensor, # Noisy Latents [B, C, H, W] or [B, C, F, H, W]
        timestep: torch.LongTensor,
        encoder_hidden_states: torch.Tensor, # Text Embeddings
        visual_conditions: Optional[List[torch.Tensor]] = None, # List of [B, L, D]
        video_frames: Optional[int] = None, # If generating video
        return_dict: bool = True,
    ) -> Union[torch.Tensor, BaseOutput]:
        
        batch_size, channels, _, _ = hidden_states.shape
        
        # 1. Patchify Logic (supports video 3D patching implicitly if reshaped)
        # Simplified for 2D view: [B, C, H, W] -> [B, (H/P * W/P), C*P*P]
        p = self.config.patch_size
        h, w = hidden_states.shape[-2], hidden_states.shape[-1]
        x = hidden_states.unfold(2, p, p).unfold(3, p, p)
        x = x.permute(0, 2, 3, 1, 4, 5).contiguous()
        x = x.view(batch_size, -1, channels * p * p) # [B, L, D_in]
        
        # 2. Embedding
        x = self.x_embedder(x)
        x = x + self.pos_embed[:, :x.shape[1], :]
        
        t = self.t_embedder(timestep, x.dtype)
        
        # 3. Process Visual Conditions (1-3 images)
        visual_emb = None
        if visual_conditions is not None:
            # Stack and project: expect list of tensors
            # Professional handling: Concatenate along sequence dim
            concat_visuals = torch.cat(visual_conditions, dim=1) # [B, Total_L, Vis_Dim]
            visual_emb = self.visual_projector(concat_visuals)

        # 4. Transformer Blocks
        for block in self.blocks:
            x = block(
                hidden_states=x, 
                encoder_hidden_states=encoder_hidden_states,
                visual_context=visual_emb,
                timestep_emb=t
            )

        # 5. Output Projector
        x = self.final_layer[0](x) # Norm
        
        # AdaLN shift/scale for final layer (simplified from DiT paper)
        # x = x * (1 + scale) + shift ... omitted for brevity
        
        x = self.final_layer[1](x) # Linear projection
        
        # 6. Unpatchify
        output = self.unpatchify(x, h, w)

        if not return_dict:
            return (output,)

        return BaseOutput(sample=output)

# -----------------------------------------------------------------------------
# 5. The "Fancy" Pipeline
# -----------------------------------------------------------------------------

class OmniMMDitV2Pipeline(DiffusionPipeline):
    """
    Pipeline for Omni-Modal Image/Video Editing.
    Features:
    - Multi-modal conditioning (Text + Multi-Image)
    - Video generation support
    - Fancy progress bar and callback support
    """
    model: OmniMMDitV2
    tokenizer: CLIPTokenizer
    text_encoder: CLIPTextModel
    vae: Any # AutoencoderKL
    scheduler: DDIMScheduler
    
    _optional_components = ["visual_encoder"]

    def __init__(
        self,
        model: OmniMMDitV2,
        vae: Any,
        text_encoder: CLIPTextModel,
        tokenizer: CLIPTokenizer,
        scheduler: DDIMScheduler,
        visual_encoder: Optional[Any] = None,
    ):
        super().__init__()
        self.register_modules(
            model=model,
            vae=vae,
            text_encoder=text_encoder,
            tokenizer=tokenizer,
            scheduler=scheduler,
            visual_encoder=visual_encoder
        )
        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)

    @torch.no_grad()
    def __call__(
        self,
        prompt: Union[str, List[str]] = None,
        input_images: Optional[List[Union[torch.Tensor, Any]]] = None, # 1-3 Images
        height: Optional[int] = 1024,
        width: Optional[int] = 1024,
        num_frames: Optional[int] = 1, # >1 triggers video mode
        num_inference_steps: int = 50,
        guidance_scale: float = 7.5,
        image_guidance_scale: float = 1.5,
        negative_prompt: Optional[Union[str, List[str]]] = None,
        eta: float = 0.0,
        generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
        latents: Optional[torch.Tensor] = None,
        output_type: Optional[str] = "pil",
        return_dict: bool = True,
        **kwargs,
    ):
        # 0. Default height/width
        height = height or self.model.config.sample_size * self.vae_scale_factor
        width = width or self.model.config.sample_size * self.vae_scale_factor

        # 1. Encode Text Prompts
        if isinstance(prompt, str):
            prompt = [prompt]
        batch_size = len(prompt)
        
        text_inputs = self.tokenizer(
            prompt, padding="max_length", max_length=self.tokenizer.model_max_length, truncation=True, return_tensors="pt"
        )
        text_embeddings = self.text_encoder(text_inputs.input_ids.to(self.device))[0]

        # 2. Encode Visual Conditions (Complex Fancy Logic)
        visual_embeddings_list = []
        if input_images:
            if not isinstance(input_images, list):
                input_images = [input_images]
            if len(input_images) > 3:
                raise ValueError("OmniMMDitV2 supports a maximum of 3 reference images.")
            
            # Simulate Visual Encoder (e.g. CLIP Vision)
            for img in input_images:
                # In real pipeline: preprocess -> visual_encoder -> project
                # Here we simulate the embedding for structural completeness
                dummy_vis = torch.randn((batch_size, 257, self.model.config.visual_embed_dim), device=self.device, dtype=text_embeddings.dtype)
                visual_embeddings_list.append(dummy_vis)

        # 3. Prepare Timesteps
        self.scheduler.set_timesteps(num_inference_steps, device=self.device)
        timesteps = self.scheduler.timesteps

        # 4. Prepare Latents (Noise)
        num_channels_latents = self.model.config.in_channels
        shape = (batch_size, num_channels_latents, height // self.vae_scale_factor, width // self.vae_scale_factor)
        
        # Handle Video Latents (5D)
        if num_frames > 1:
            shape = (batch_size, num_channels_latents, num_frames, height // self.vae_scale_factor, width // self.vae_scale_factor)

        latents = torch.randn(shape, generator=generator, device=self.device, dtype=text_embeddings.dtype)
        latents = latents * self.scheduler.init_noise_sigma

        # 5. Denoising Loop (The "Fancy" Part)
        with self.progress_bar(total=num_inference_steps) as progress_bar:
            for i, t in enumerate(timesteps):
                # Expand latents for classifier-free guidance
                latent_model_input = torch.cat([latents] * 2) if guidance_scale > 1.0 else latents
                latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)

                # Predict noise
                # Handle Classifier Free Guidance (Text + Image)
                # We duplicate text embeddings for unconditional pass (usually empty string)
                # Omitted complex CFG setup for brevity, assuming simple split
                
                noise_pred = self.model(
                    hidden_states=latent_model_input,
                    timestep=t,
                    encoder_hidden_states=torch.cat([text_embeddings] * 2), # Simplified
                    visual_conditions=visual_embeddings_list * 2 if visual_embeddings_list else None,
                    video_frames=num_frames
                ).sample

                # Perform Guidance
                if guidance_scale > 1.0:
                    noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
                    noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)

                # Compute previous noisy sample x_t -> x_t-1
                latents = self.scheduler.step(noise_pred, t, latents, eta=eta).prev_sample
                progress_bar.update()

        # 6. Post-processing
        if not output_type == "latent":
            # self.vae.decode(latents / self.vae.config.scaling_factor) ...
            pass # VAE Decode Logic

        if not return_dict:
            return (latents,)

        return BaseOutput(images=latents) # Returning latents for simulation