V1.5.0

diffusers_helper/bucket_tools.py

@@ -0,0 +1,36 @@
+bucket_options = {
+    (416, 960),
+    (448, 864),
+    (480, 832),
+    (512, 768),
+    (544, 704),
+    (576, 672),
+    (608, 640),
+    (640, 608),
+    (672, 576),
+    (704, 544),
+    (768, 512),
+    (832, 480),
+    (864, 448),
+    (960, 416),
+}
+
+
+def find_nearest_bucket(h, w, resolution=640):
+    min_metric = float('inf')
+    best_bucket = None
+    for (bucket_h, bucket_w) in bucket_options:
+        metric = abs(h * bucket_w - w * bucket_h)
+        if metric <= min_metric:
+            min_metric = metric
+            best_bucket = (bucket_h, bucket_w)
+
+    if resolution != 640:
+        scale_factor = resolution / 640.0
+        scaled_height = round(best_bucket[0] * scale_factor / 16) * 16
+        scaled_width = round(best_bucket[1] * scale_factor / 16) * 16
+        best_bucket = (scaled_height, scaled_width)
+        print(f'Resolution: {best_bucket[1]} x {best_bucket[0]}')
+
+    return best_bucket
+

diffusers_helper/clip_vision.py

@@ -0,0 +1,12 @@
+import numpy as np
+
+
+def hf_clip_vision_encode(image, feature_extractor, image_encoder):
+    assert isinstance(image, np.ndarray)
+    assert image.ndim == 3 and image.shape[2] == 3
+    assert image.dtype == np.uint8
+
+    preprocessed = feature_extractor.preprocess(images=image, return_tensors="pt").to(device=image_encoder.device, dtype=image_encoder.dtype)
+    image_encoder_output = image_encoder(**preprocessed)
+
+    return image_encoder_output

diffusers_helper/dit_common.py

@@ -0,0 +1,53 @@
+import torch
+import accelerate.accelerator
+
+from diffusers.models.normalization import RMSNorm, LayerNorm, FP32LayerNorm, AdaLayerNormContinuous
+
+
+accelerate.accelerator.convert_outputs_to_fp32 = lambda x: x
+
+
+def LayerNorm_forward(self, x):
+    return torch.nn.functional.layer_norm(x, self.normalized_shape, self.weight, self.bias, self.eps).to(x)
+
+
+LayerNorm.forward = LayerNorm_forward
+torch.nn.LayerNorm.forward = LayerNorm_forward
+
+
+def FP32LayerNorm_forward(self, x):
+    origin_dtype = x.dtype
+    return torch.nn.functional.layer_norm(
+        x.float(),
+        self.normalized_shape,
+        self.weight.float() if self.weight is not None else None,
+        self.bias.float() if self.bias is not None else None,
+        self.eps,
+    ).to(origin_dtype)
+
+
+FP32LayerNorm.forward = FP32LayerNorm_forward
+
+
+def RMSNorm_forward(self, hidden_states):
+    input_dtype = hidden_states.dtype
+    variance = hidden_states.to(torch.float32).pow(2).mean(-1, keepdim=True)
+    hidden_states = hidden_states * torch.rsqrt(variance + self.eps)
+
+    if self.weight is None:
+        return hidden_states.to(input_dtype)
+
+    return hidden_states.to(input_dtype) * self.weight.to(input_dtype)
+
+
+RMSNorm.forward = RMSNorm_forward
+
+
+def AdaLayerNormContinuous_forward(self, x, conditioning_embedding):
+    emb = self.linear(self.silu(conditioning_embedding))
+    scale, shift = emb.chunk(2, dim=1)
+    x = self.norm(x) * (1 + scale)[:, None, :] + shift[:, None, :]
+    return x
+
+
+AdaLayerNormContinuous.forward = AdaLayerNormContinuous_forward

diffusers_helper/hf_login.py

@@ -0,0 +1,21 @@
+import os
+
+
+def login(token):
+    from huggingface_hub import login
+    import time
+
+    while True:
+        try:
+            login(token)
+            print('HF login ok.')
+            break
+        except Exception as e:
+            print(f'HF login failed: {e}. Retrying')
+            time.sleep(0.5)
+
+
+hf_token = os.environ.get('HF_TOKEN', None)
+
+if hf_token is not None:
+    login(hf_token)

diffusers_helper/hunyuan.py

@@ -0,0 +1,109 @@
+import torch
+
+from diffusers.pipelines.hunyuan_video.pipeline_hunyuan_video import DEFAULT_PROMPT_TEMPLATE
+
+@torch.no_grad()
+def encode_prompt_conds(prompt, text_encoder, text_encoder_2, tokenizer, tokenizer_2, max_length=256):
+    assert isinstance(prompt, str)
+
+    prompt = [prompt]
+
+    # LLAMA
+
+    prompt_llama = [DEFAULT_PROMPT_TEMPLATE["template"].format(p) for p in prompt]
+    crop_start = DEFAULT_PROMPT_TEMPLATE["crop_start"]
+
+    llama_inputs = tokenizer(
+        prompt_llama,
+        padding="max_length",
+        max_length=max_length + crop_start,
+        truncation=True,
+        return_tensors="pt",
+        return_length=False,
+        return_overflowing_tokens=False,
+        return_attention_mask=True,
+    )
+
+    llama_input_ids = llama_inputs.input_ids.to(text_encoder.device)
+    llama_attention_mask = llama_inputs.attention_mask.to(text_encoder.device)
+    llama_attention_length = int(llama_attention_mask.sum())
+
+    llama_outputs = text_encoder(
+        input_ids=llama_input_ids,
+        attention_mask=llama_attention_mask,
+        output_hidden_states=True,
+    )
+
+    llama_vec = llama_outputs.hidden_states[-3][:, crop_start:llama_attention_length]
+    # llama_vec_remaining = llama_outputs.hidden_states[-3][:, llama_attention_length:]
+    llama_attention_mask = llama_attention_mask[:, crop_start:llama_attention_length]
+
+    assert torch.all(llama_attention_mask.bool())
+
+    # CLIP
+
+    clip_l_input_ids = tokenizer_2(
+        prompt,
+        padding="max_length",
+        max_length=77,
+        truncation=True,
+        return_overflowing_tokens=False,
+        return_length=False,
+        return_tensors="pt",
+    ).input_ids
+    clip_l_pooler = text_encoder_2(clip_l_input_ids.to(text_encoder_2.device), output_hidden_states=False).pooler_output
+
+    return llama_vec, clip_l_pooler
+
+
+@torch.no_grad()
+def vae_decode_fake(latents):
+    latent_rgb_factors = [
+        [-0.0395, -0.0331, 0.0445],
+        [0.0696, 0.0795, 0.0518],
+        [0.0135, -0.0945, -0.0282],
+        [0.0108, -0.0250, -0.0765],
+        [-0.0209, 0.0032, 0.0224],
+        [-0.0804, -0.0254, -0.0639],
+        [-0.0991, 0.0271, -0.0669],
+        [-0.0646, -0.0422, -0.0400],
+        [-0.0696, -0.0595, -0.0894],
+        [-0.0799, -0.0208, -0.0375],
+        [0.1166, 0.1627, 0.0962],
+        [0.1165, 0.0432, 0.0407],
+        [-0.2315, -0.1920, -0.1355],
+        [-0.0270, 0.0401, -0.0821],
+        [-0.0616, -0.0997, -0.0727],
+        [0.0249, -0.0469, -0.1703]
+    ]  # From comfyui
+
+    latent_rgb_factors_bias = [0.0259, -0.0192, -0.0761]
+
+    weight = torch.tensor(latent_rgb_factors, device=latents.device, dtype=latents.dtype).transpose(0, 1)[:, :, None, None, None]
+    bias = torch.tensor(latent_rgb_factors_bias, device=latents.device, dtype=latents.dtype)
+
+    images = torch.nn.functional.conv3d(latents, weight, bias=bias, stride=1, padding=0, dilation=1, groups=1)
+    images = images.clamp(0.0, 1.0)
+
+    return images
+
+
+@torch.no_grad()
+def vae_decode(latents, vae, image_mode=False):
+    latents = latents / vae.config.scaling_factor
+
+    if not image_mode:
+        image = vae.decode(latents.to(device=vae.device, dtype=vae.dtype)).sample
+    else:
+        latents = latents.to(device=vae.device, dtype=vae.dtype).unbind(2)
+        image = [vae.decode(l.unsqueeze(2)).sample for l in latents]
+        image = torch.cat(image, dim=2)
+
+    return image
+
+
+@torch.no_grad()
+def vae_encode(image, vae):
+    latents = vae.encode(image.to(device=vae.device, dtype=vae.dtype)).latent_dist.sample()
+    latents = latents * vae.config.scaling_factor
+    return latents

diffusers_helper/k_diffusion_hunyuan.py

@@ -0,0 +1,221 @@
+import torch
+import math
+
+from .k_diffusion.uni_pc_fm import sample_unipc
+from .k_diffusion.wrapper import fm_wrapper
+from .utils import repeat_to_batch_size
+
+
+def flux_time_shift(t, mu=1.15, sigma=1.0):
+    return math.exp(mu) / (math.exp(mu) + (1 / t - 1) ** sigma)
+
+
+def calculate_flux_mu(context_length, x1=256, y1=0.5, x2=4096, y2=1.15, exp_max=7.0):
+    k = (y2 - y1) / (x2 - x1)
+    b = y1 - k * x1
+    mu = k * context_length + b
+    mu = min(mu, math.log(exp_max))
+    return mu
+
+
+def get_flux_sigmas_from_mu(n, mu):
+    sigmas = torch.linspace(1, 0, steps=n + 1)
+    sigmas = flux_time_shift(sigmas, mu=mu)
+    return sigmas
+
+
+@torch.inference_mode()
+def sample_hunyuan2(
+        transformer,
+        sampler='unipc',
+        initial_latent=None,
+        concat_latent=None,
+        strength=1.0,
+        width=512,
+        height=512,
+        frames=16,
+        real_guidance_scale=1.0,
+        distilled_guidance_scale=6.0,
+        guidance_rescale=0.0,
+        shift=None,
+        num_inference_steps=25,
+        batch_size=None,
+        generator=None,
+        noise=None,
+        prompt_embeds=None,
+        prompt_embeds_mask=None,
+        prompt_poolers=None,
+        negative_prompt_embeds=None,
+        negative_prompt_embeds_mask=None,
+        negative_prompt_poolers=None,
+        dtype=torch.bfloat16,
+        device=None,
+        negative_kwargs=None,
+        callback=None,
+        **kwargs,
+):
+    device = device or transformer.device
+
+    if batch_size is None:
+        batch_size = int(prompt_embeds.shape[0])
+
+    if noise is None:
+        latents = torch.randn((batch_size, 16, (frames + 3) // 4, height // 8, width // 8), generator=generator, device=generator.device).to(device=device, dtype=torch.float32)
+    else:
+        latents = noise.clone().detach().to(device=device, dtype=torch.float32)
+
+    B, C, T, H, W = latents.shape
+    seq_length = T * H * W // 4
+
+    if shift is None:
+        mu = calculate_flux_mu(seq_length, exp_max=7.0)
+    else:
+        mu = math.log(shift)
+
+    sigmas = get_flux_sigmas_from_mu(num_inference_steps, mu).to(device)
+
+    k_model = fm_wrapper(transformer)
+
+    if initial_latent is not None:
+        sigmas = sigmas * strength
+        first_sigma = sigmas[0].to(device=device, dtype=torch.float32)
+        initial_latent = initial_latent.to(device=device, dtype=torch.float32)
+        latents = initial_latent.float() * (1.0 - first_sigma) + latents.float() * first_sigma
+
+    if concat_latent is not None:
+        concat_latent = concat_latent.to(latents)
+
+    distilled_guidance = torch.tensor([distilled_guidance_scale * 1000.0] * batch_size).to(device=device, dtype=dtype)
+
+    prompt_embeds = repeat_to_batch_size(prompt_embeds, batch_size)
+    prompt_embeds_mask = repeat_to_batch_size(prompt_embeds_mask, batch_size)
+    prompt_poolers = repeat_to_batch_size(prompt_poolers, batch_size)
+    negative_prompt_embeds = repeat_to_batch_size(negative_prompt_embeds, batch_size)
+    negative_prompt_embeds_mask = repeat_to_batch_size(negative_prompt_embeds_mask, batch_size)
+    negative_prompt_poolers = repeat_to_batch_size(negative_prompt_poolers, batch_size)
+    concat_latent = repeat_to_batch_size(concat_latent, batch_size)
+
+    sampler_kwargs = dict(
+        dtype=dtype,
+        cfg_scale=real_guidance_scale,
+        cfg_rescale=guidance_rescale,
+        concat_latent=concat_latent,
+        positive=dict(
+            pooled_projections=prompt_poolers,
+            encoder_hidden_states=prompt_embeds,
+            encoder_attention_mask=prompt_embeds_mask,
+            guidance=distilled_guidance,
+            **kwargs,
+        ),
+        negative=dict(
+            pooled_projections=negative_prompt_poolers,
+            encoder_hidden_states=negative_prompt_embeds,
+            encoder_attention_mask=negative_prompt_embeds_mask,
+            guidance=distilled_guidance,
+            **(kwargs if negative_kwargs is None else {**kwargs, **negative_kwargs}),
+        )
+    )
+
+    if sampler == 'unipc_bh1':
+        variant = 'bh1'
+    elif sampler == 'unipc_bh2':
+        variant = 'bh2'
+    results = sample_unipc(k_model, latents, sigmas, extra_args=sampler_kwargs, disable=False, variant=variant, callback=callback)
+
+    return results
+
+@torch.inference_mode()
+def sample_hunyuan(
+        transformer,
+        sampler='unipc',
+        initial_latent=None,
+        concat_latent=None,
+        strength=1.0,
+        width=512,
+        height=512,
+        frames=16,
+        real_guidance_scale=1.0,
+        distilled_guidance_scale=6.0,
+        guidance_rescale=0.0,
+        shift=None,
+        num_inference_steps=25,
+        batch_size=None,
+        generator=None,
+        prompt_embeds=None,
+        prompt_embeds_mask=None,
+        prompt_poolers=None,
+        negative_prompt_embeds=None,
+        negative_prompt_embeds_mask=None,
+        negative_prompt_poolers=None,
+        dtype=torch.bfloat16,
+        device=None,
+        negative_kwargs=None,
+        callback=None,
+        **kwargs,
+):
+    device = device or transformer.device
+
+    if batch_size is None:
+        batch_size = int(prompt_embeds.shape[0])
+
+    latents = torch.randn((batch_size, 16, (frames + 3) // 4, height // 8, width // 8), generator=generator, device=generator.device).to(device=device, dtype=torch.float32)
+
+    B, C, T, H, W = latents.shape
+    seq_length = T * H * W // 4
+
+    if shift is None:
+        mu = calculate_flux_mu(seq_length, exp_max=7.0)
+    else:
+        mu = math.log(shift)
+
+    sigmas = get_flux_sigmas_from_mu(num_inference_steps, mu).to(device)
+
+    k_model = fm_wrapper(transformer)
+
+    if initial_latent is not None:
+        sigmas = sigmas * strength
+        first_sigma = sigmas[0].to(device=device, dtype=torch.float32)
+        initial_latent = initial_latent.to(device=device, dtype=torch.float32)
+        latents = initial_latent.float() * (1.0 - first_sigma) + latents.float() * first_sigma
+
+    if concat_latent is not None:
+        concat_latent = concat_latent.to(latents)
+
+    distilled_guidance = torch.tensor([distilled_guidance_scale * 1000.0] * batch_size).to(device=device, dtype=dtype)
+
+    prompt_embeds = repeat_to_batch_size(prompt_embeds, batch_size)
+    prompt_embeds_mask = repeat_to_batch_size(prompt_embeds_mask, batch_size)
+    prompt_poolers = repeat_to_batch_size(prompt_poolers, batch_size)
+    negative_prompt_embeds = repeat_to_batch_size(negative_prompt_embeds, batch_size)
+    negative_prompt_embeds_mask = repeat_to_batch_size(negative_prompt_embeds_mask, batch_size)
+    negative_prompt_poolers = repeat_to_batch_size(negative_prompt_poolers, batch_size)
+    concat_latent = repeat_to_batch_size(concat_latent, batch_size)
+
+    sampler_kwargs = dict(
+        dtype=dtype,
+        cfg_scale=real_guidance_scale,
+        cfg_rescale=guidance_rescale,
+        concat_latent=concat_latent,
+        positive=dict(
+            pooled_projections=prompt_poolers,
+            encoder_hidden_states=prompt_embeds,
+            encoder_attention_mask=prompt_embeds_mask,
+            guidance=distilled_guidance,
+            **kwargs,
+        ),
+        negative=dict(
+            pooled_projections=negative_prompt_poolers,
+            encoder_hidden_states=negative_prompt_embeds,
+            encoder_attention_mask=negative_prompt_embeds_mask,
+            guidance=distilled_guidance,
+            **(kwargs if negative_kwargs is None else {**kwargs, **negative_kwargs}),
+        )
+    )
+
+    if sampler == 'unipc_bh1':
+        variant = 'bh1'
+    elif sampler == 'unipc_bh2':
+        variant = 'bh2'
+    results = sample_unipc(k_model, latents, sigmas, extra_args=sampler_kwargs, disable=False, variant=variant, callback=callback)
+
+    return results

diffusers_helper/memory.py

@@ -0,0 +1,134 @@
+# By lllyasviel
+
+
+import torch
+
+
+cpu = torch.device('cpu')
+gpu = torch.device(f'cuda:{torch.cuda.current_device()}')
+gpu_complete_modules = []
+
+
+class DynamicSwapInstaller:
+    @staticmethod
+    def _install_module(module: torch.nn.Module, **kwargs):
+        original_class = module.__class__
+        module.__dict__['forge_backup_original_class'] = original_class
+
+        def hacked_get_attr(self, name: str):
+            if '_parameters' in self.__dict__:
+                _parameters = self.__dict__['_parameters']
+                if name in _parameters:
+                    p = _parameters[name]
+                    if p is None:
+                        return None
+                    if p.__class__ == torch.nn.Parameter:
+                        return torch.nn.Parameter(p.to(**kwargs), requires_grad=p.requires_grad)
+                    else:
+                        return p.to(**kwargs)
+            if '_buffers' in self.__dict__:
+                _buffers = self.__dict__['_buffers']
+                if name in _buffers:
+                    return _buffers[name].to(**kwargs)
+            return super(original_class, self).__getattr__(name)
+
+        module.__class__ = type('DynamicSwap_' + original_class.__name__, (original_class,), {
+            '__getattr__': hacked_get_attr,
+        })
+
+        return
+
+    @staticmethod
+    def _uninstall_module(module: torch.nn.Module):
+        if 'forge_backup_original_class' in module.__dict__:
+            module.__class__ = module.__dict__.pop('forge_backup_original_class')
+        return
+
+    @staticmethod
+    def install_model(model: torch.nn.Module, **kwargs):
+        for m in model.modules():
+            DynamicSwapInstaller._install_module(m, **kwargs)
+        return
+
+    @staticmethod
+    def uninstall_model(model: torch.nn.Module):
+        for m in model.modules():
+            DynamicSwapInstaller._uninstall_module(m)
+        return
+
+
+def fake_diffusers_current_device(model: torch.nn.Module, target_device: torch.device):
+    if hasattr(model, 'scale_shift_table'):
+        model.scale_shift_table.data = model.scale_shift_table.data.to(target_device)
+        return
+
+    for k, p in model.named_modules():
+        if hasattr(p, 'weight'):
+            p.to(target_device)
+            return
+
+
+def get_cuda_free_memory_gb(device=None):
+    if device is None:
+        device = gpu
+
+    memory_stats = torch.cuda.memory_stats(device)
+    bytes_active = memory_stats['active_bytes.all.current']
+    bytes_reserved = memory_stats['reserved_bytes.all.current']
+    bytes_free_cuda, _ = torch.cuda.mem_get_info(device)
+    bytes_inactive_reserved = bytes_reserved - bytes_active
+    bytes_total_available = bytes_free_cuda + bytes_inactive_reserved
+    return bytes_total_available / (1024 ** 3)
+
+
+def move_model_to_device_with_memory_preservation(model, target_device, preserved_memory_gb=0):
+    print(f'Moving {model.__class__.__name__} to {target_device} with preserved memory: {preserved_memory_gb} GB')
+
+    for m in model.modules():
+        if get_cuda_free_memory_gb(target_device) <= preserved_memory_gb:
+            torch.cuda.empty_cache()
+            return
+
+        if hasattr(m, 'weight'):
+            m.to(device=target_device)
+
+    model.to(device=target_device)
+    torch.cuda.empty_cache()
+    return
+
+
+def offload_model_from_device_for_memory_preservation(model, target_device, preserved_memory_gb=0):
+    print(f'Offloading {model.__class__.__name__} from {target_device} to preserve memory: {preserved_memory_gb} GB')
+
+    for m in model.modules():
+        if get_cuda_free_memory_gb(target_device) >= preserved_memory_gb:
+            torch.cuda.empty_cache()
+            return
+
+        if hasattr(m, 'weight'):
+            m.to(device=cpu)
+
+    model.to(device=cpu)
+    torch.cuda.empty_cache()
+    return
+
+
+def unload_complete_models(*args):
+    for m in gpu_complete_modules + list(args):
+        m.to(device=cpu)
+        print(f'Unloaded {m.__class__.__name__} as complete.')
+
+    gpu_complete_modules.clear()
+    torch.cuda.empty_cache()
+    return
+
+
+def load_model_as_complete(model, target_device, unload=True):
+    if unload:
+        unload_complete_models()
+
+    model.to(device=target_device)
+    print(f'Loaded {model.__class__.__name__} to {target_device} as complete.')
+
+    gpu_complete_modules.append(model)
+    return

diffusers_helper/thread_utils.py

@@ -0,0 +1,76 @@
+import time
+
+from threading import Thread, Lock
+
+
+class Listener:
+    task_queue = []
+    lock = Lock()
+    thread = None
+    
+    @classmethod
+    def _process_tasks(cls):
+        while True:
+            task = None
+            with cls.lock:
+                if cls.task_queue:
+                    task = cls.task_queue.pop(0)
+                    
+            if task is None:
+                time.sleep(0.001)
+                continue
+                
+            func, args, kwargs = task
+            try:
+                func(*args, **kwargs)
+            except Exception as e:
+                print(f"Error in listener thread: {e}")
+    
+    @classmethod
+    def add_task(cls, func, *args, **kwargs):
+        with cls.lock:
+            cls.task_queue.append((func, args, kwargs))
+
+        if cls.thread is None:
+            cls.thread = Thread(target=cls._process_tasks, daemon=True)
+            cls.thread.start()
+
+
+def async_run(func, *args, **kwargs):
+    Listener.add_task(func, *args, **kwargs)
+
+
+class FIFOQueue:
+    def __init__(self):
+        self.queue = []
+        self.lock = Lock()
+
+    def push(self, item):
+        with self.lock:
+            self.queue.append(item)
+
+    def pop(self):
+        with self.lock:
+            if self.queue:
+                return self.queue.pop(0)
+            return None
+
+    def top(self):
+        with self.lock:
+            if self.queue:
+                return self.queue[0]
+            return None
+
+    def next(self):
+        while True:
+            with self.lock:
+                if self.queue:
+                    return self.queue.pop(0)
+
+            time.sleep(0.001)
+
+
+class AsyncStream:
+    def __init__(self):
+        self.input_queue = FIFOQueue()
+        self.output_queue = FIFOQueue()

diffusers_helper/utils.py

@@ -0,0 +1,622 @@
+import os
+import cv2
+import json
+import random
+import glob
+import torch
+import einops
+import numpy as np
+import datetime
+import torchvision
+
+import safetensors.torch as sf
+from PIL import Image
+
+
+def min_resize(x, m):
+    if x.shape[0] < x.shape[1]:
+        s0 = m
+        s1 = int(float(m) / float(x.shape[0]) * float(x.shape[1]))
+    else:
+        s0 = int(float(m) / float(x.shape[1]) * float(x.shape[0]))
+        s1 = m
+    new_max = max(s1, s0)
+    raw_max = max(x.shape[0], x.shape[1])
+    if new_max < raw_max:
+        interpolation = cv2.INTER_AREA
+    else:
+        interpolation = cv2.INTER_LANCZOS4
+    y = cv2.resize(x, (s1, s0), interpolation=interpolation)
+    return y
+
+
+def d_resize(x, y):
+    H, W, C = y.shape
+    new_min = min(H, W)
+    raw_min = min(x.shape[0], x.shape[1])
+    if new_min < raw_min:
+        interpolation = cv2.INTER_AREA
+    else:
+        interpolation = cv2.INTER_LANCZOS4
+    y = cv2.resize(x, (W, H), interpolation=interpolation)
+    return y
+
+
+def resize_and_center_crop(image, target_width, target_height):
+    if target_height == image.shape[0] and target_width == image.shape[1]:
+        return image
+
+    pil_image = Image.fromarray(image)
+    original_width, original_height = pil_image.size
+    scale_factor = max(target_width / original_width, target_height / original_height)
+    resized_width = int(round(original_width * scale_factor))
+    resized_height = int(round(original_height * scale_factor))
+    resized_image = pil_image.resize((resized_width, resized_height), Image.LANCZOS)
+    left = (resized_width - target_width) / 2
+    top = (resized_height - target_height) / 2
+    right = (resized_width + target_width) / 2
+    bottom = (resized_height + target_height) / 2
+    cropped_image = resized_image.crop((left, top, right, bottom))
+    return np.array(cropped_image)
+
+
+def resize_and_center_crop_pytorch(image, target_width, target_height):
+    B, C, H, W = image.shape
+
+    if H == target_height and W == target_width:
+        return image
+
+    scale_factor = max(target_width / W, target_height / H)
+    resized_width = int(round(W * scale_factor))
+    resized_height = int(round(H * scale_factor))
+
+    resized = torch.nn.functional.interpolate(image, size=(resized_height, resized_width), mode='bilinear', align_corners=False)
+
+    top = (resized_height - target_height) // 2
+    left = (resized_width - target_width) // 2
+    cropped = resized[:, :, top:top + target_height, left:left + target_width]
+
+    return cropped
+
+
+def resize_without_crop(image, target_width, target_height):
+    if target_height == image.shape[0] and target_width == image.shape[1]:
+        return image
+
+    pil_image = Image.fromarray(image)
+    resized_image = pil_image.resize((target_width, target_height), Image.LANCZOS)
+    return np.array(resized_image)
+
+
+def just_crop(image, w, h):
+    if h == image.shape[0] and w == image.shape[1]:
+        return image
+
+    original_height, original_width = image.shape[:2]
+    k = min(original_height / h, original_width / w)
+    new_width = int(round(w * k))
+    new_height = int(round(h * k))
+    x_start = (original_width - new_width) // 2
+    y_start = (original_height - new_height) // 2
+    cropped_image = image[y_start:y_start + new_height, x_start:x_start + new_width]
+    return cropped_image
+
+
+def write_to_json(data, file_path):
+    temp_file_path = file_path + ".tmp"
+    with open(temp_file_path, 'wt', encoding='utf-8') as temp_file:
+        json.dump(data, temp_file, indent=4)
+    os.replace(temp_file_path, file_path)
+    return
+
+
+def read_from_json(file_path):
+    with open(file_path, 'rt', encoding='utf-8') as file:
+        data = json.load(file)
+    return data
+
+
+def get_active_parameters(m):
+    return {k: v for k, v in m.named_parameters() if v.requires_grad}
+
+
+def cast_training_params(m, dtype=torch.float32):
+    result = {}
+    for n, param in m.named_parameters():
+        if param.requires_grad:
+            param.data = param.to(dtype)
+            result[n] = param
+    return result
+
+
+def separate_lora_AB(parameters, B_patterns=None):
+    parameters_normal = {}
+    parameters_B = {}
+
+    if B_patterns is None:
+        B_patterns = ['.lora_B.', '__zero__']
+
+    for k, v in parameters.items():
+        if any(B_pattern in k for B_pattern in B_patterns):
+            parameters_B[k] = v
+        else:
+            parameters_normal[k] = v
+
+    return parameters_normal, parameters_B
+
+
+def set_attr_recursive(obj, attr, value):
+    attrs = attr.split(".")
+    for name in attrs[:-1]:
+        obj = getattr(obj, name)
+    setattr(obj, attrs[-1], value)
+    return
+
+
+def print_tensor_list_size(tensors):
+    total_size = 0
+    total_elements = 0
+
+    if isinstance(tensors, dict):
+        tensors = tensors.values()
+
+    for tensor in tensors:
+        total_size += tensor.nelement() * tensor.element_size()
+        total_elements += tensor.nelement()
+
+    total_size_MB = total_size / (1024 ** 2)
+    total_elements_B = total_elements / 1e9
+
+    print(f"Total number of tensors: {len(tensors)}")
+    print(f"Total size of tensors: {total_size_MB:.2f} MB")
+    print(f"Total number of parameters: {total_elements_B:.3f} billion")
+    return
+
+
+@torch.no_grad()
+def batch_mixture(a, b=None, probability_a=0.5, mask_a=None):
+    batch_size = a.size(0)
+
+    if b is None:
+        b = torch.zeros_like(a)
+
+    if mask_a is None:
+        mask_a = torch.rand(batch_size) < probability_a
+
+    mask_a = mask_a.to(a.device)
+    mask_a = mask_a.reshape((batch_size,) + (1,) * (a.dim() - 1))
+    result = torch.where(mask_a, a, b)
+    return result
+
+
+@torch.no_grad()
+def zero_module(module):
+    for p in module.parameters():
+        p.detach().zero_()
+    return module
+
+
+@torch.no_grad()
+def supress_lower_channels(m, k, alpha=0.01):
+    data = m.weight.data.clone()
+
+    assert int(data.shape[1]) >= k
+
+    data[:, :k] = data[:, :k] * alpha
+    m.weight.data = data.contiguous().clone()
+    return m
+
+
+def freeze_module(m):
+    if not hasattr(m, '_forward_inside_frozen_module'):
+        m._forward_inside_frozen_module = m.forward
+    m.requires_grad_(False)
+    m.forward = torch.no_grad()(m.forward)
+    return m
+
+
+def get_latest_safetensors(folder_path):
+    safetensors_files = glob.glob(os.path.join(folder_path, '*.safetensors'))
+
+    if not safetensors_files:
+        raise ValueError('No file to resume!')
+
+    latest_file = max(safetensors_files, key=os.path.getmtime)
+    latest_file = os.path.abspath(os.path.realpath(latest_file))
+    return latest_file
+
+
+def generate_random_prompt_from_tags(tags_str, min_length=3, max_length=32):
+    tags = tags_str.split(', ')
+    tags = random.sample(tags, k=min(random.randint(min_length, max_length), len(tags)))
+    prompt = ', '.join(tags)
+    return prompt
+
+
+def interpolate_numbers(a, b, n, round_to_int=False, gamma=1.0):
+    numbers = a + (b - a) * (np.linspace(0, 1, n) ** gamma)
+    if round_to_int:
+        numbers = np.round(numbers).astype(int)
+    return numbers.tolist()
+
+
+def uniform_random_by_intervals(inclusive, exclusive, n, round_to_int=False):
+    edges = np.linspace(0, 1, n + 1)
+    points = np.random.uniform(edges[:-1], edges[1:])
+    numbers = inclusive + (exclusive - inclusive) * points
+    if round_to_int:
+        numbers = np.round(numbers).astype(int)
+    return numbers.tolist()
+
+
+def soft_append_bcthw(history, current, overlap=0):
+    if overlap <= 0:
+        return torch.cat([history, current], dim=2)
+
+    assert history.shape[2] >= overlap, f"History length ({history.shape[2]}) must be >= overlap ({overlap})"
+    assert current.shape[2] >= overlap, f"Current length ({current.shape[2]}) must be >= overlap ({overlap})"
+
+    print("soft_append_bcthw: history Shape: {}".format(history.shape))
+    print("soft_append_bcthw: history Shape: {}".format(current.shape))
+    print("soft_append_bcthw: overlap: {}".format(overlap))
+
+    weights = torch.linspace(1, 0, overlap, dtype=history.dtype, device=history.device).view(1, 1, -1, 1, 1)
+    
+    print("weights: Shape: {}".format(weights.shape))
+    print("weights: history[:, :, -overlap:] Shape: {}".format(history[:, :, -overlap:].shape))
+    print("weights: current[:, :, overlap:] Shape: {}".format(current[:, :, overlap:].shape))
+
+    blended = weights * history[:, :, -overlap:] + (1 - weights) * current[:, :, :overlap]
+    output = torch.cat([history[:, :, :-overlap], blended, current[:, :, overlap:]], dim=2)
+
+    return output.to(history)
+
+
+def save_bcthw_as_mp4(x, output_filename, fps=10):
+    b, c, t, h, w = x.shape
+
+    per_row = b
+    for p in [6, 5, 4, 3, 2]:
+        if b % p == 0:
+            per_row = p
+            break
+
+    os.makedirs(os.path.dirname(os.path.abspath(os.path.realpath(output_filename))), exist_ok=True)
+    x = torch.clamp(x.float(), -1., 1.) * 127.5 + 127.5
+    x = x.detach().cpu().to(torch.uint8)
+    x = einops.rearrange(x, '(m n) c t h w -> t (m h) (n w) c', n=per_row)
+    torchvision.io.write_video(output_filename, x, fps=fps, video_codec='h264', options={'crf': '0'})
+    return x
+
+
+def save_bcthw_as_png(x, output_filename):
+    os.makedirs(os.path.dirname(os.path.abspath(os.path.realpath(output_filename))), exist_ok=True)
+    x = torch.clamp(x.float(), -1., 1.) * 127.5 + 127.5
+    x = x.detach().cpu().to(torch.uint8)
+    x = einops.rearrange(x, 'b c t h w -> c (b h) (t w)')
+    torchvision.io.write_png(x, output_filename)
+    return output_filename
+
+
+def save_bchw_as_png(x, output_filename):
+    os.makedirs(os.path.dirname(os.path.abspath(os.path.realpath(output_filename))), exist_ok=True)
+    x = torch.clamp(x.float(), -1., 1.) * 127.5 + 127.5
+    x = x.detach().cpu().to(torch.uint8)
+    x = einops.rearrange(x, 'b c h w -> c h (b w)')
+    torchvision.io.write_png(x, output_filename)
+    return output_filename
+
+
+def add_tensors_with_padding(tensor1, tensor2):
+    if tensor1.shape == tensor2.shape:
+        return tensor1 + tensor2
+
+    shape1 = tensor1.shape
+    shape2 = tensor2.shape
+
+    new_shape = tuple(max(s1, s2) for s1, s2 in zip(shape1, shape2))
+
+    padded_tensor1 = torch.zeros(new_shape)
+    padded_tensor2 = torch.zeros(new_shape)
+
+    padded_tensor1[tuple(slice(0, s) for s in shape1)] = tensor1
+    padded_tensor2[tuple(slice(0, s) for s in shape2)] = tensor2
+
+    result = padded_tensor1 + padded_tensor2
+    return result
+
+
+def print_free_mem():
+    torch.cuda.empty_cache()
+    free_mem, total_mem = torch.cuda.mem_get_info(0)
+    free_mem_mb = free_mem / (1024 ** 2)
+    total_mem_mb = total_mem / (1024 ** 2)
+    print(f"Free memory: {free_mem_mb:.2f} MB")
+    print(f"Total memory: {total_mem_mb:.2f} MB")
+    return
+
+
+def print_gpu_parameters(device, state_dict, log_count=1):
+    summary = {"device": device, "keys_count": len(state_dict)}
+
+    logged_params = {}
+    for i, (key, tensor) in enumerate(state_dict.items()):
+        if i >= log_count:
+            break
+        logged_params[key] = tensor.flatten()[:3].tolist()
+
+    summary["params"] = logged_params
+
+    print(str(summary))
+    return
+
+
+def visualize_txt_as_img(width, height, text, font_path='font/DejaVuSans.ttf', size=18):
+    from PIL import Image, ImageDraw, ImageFont
+
+    txt = Image.new("RGB", (width, height), color="white")
+    draw = ImageDraw.Draw(txt)
+    font = ImageFont.truetype(font_path, size=size)
+
+    if text == '':
+        return np.array(txt)
+
+    # Split text into lines that fit within the image width
+    lines = []
+    words = text.split()
+    current_line = words[0]
+
+    for word in words[1:]:
+        line_with_word = f"{current_line} {word}"
+        if draw.textbbox((0, 0), line_with_word, font=font)[2] <= width:
+            current_line = line_with_word
+        else:
+            lines.append(current_line)
+            current_line = word
+
+    lines.append(current_line)
+
+    # Draw the text line by line
+    y = 0
+    line_height = draw.textbbox((0, 0), "A", font=font)[3]
+
+    for line in lines:
+        if y + line_height > height:
+            break  # stop drawing if the next line will be outside the image
+        draw.text((0, y), line, fill="black", font=font)
+        y += line_height
+
+    return np.array(txt)
+
+
+def blue_mark(x):
+    x = x.copy()
+    c = x[:, :, 2]
+    b = cv2.blur(c, (9, 9))
+    x[:, :, 2] = ((c - b) * 16.0 + b).clip(-1, 1)
+    return x
+
+
+def green_mark(x):
+    x = x.copy()
+    x[:, :, 2] = -1
+    x[:, :, 0] = -1
+    return x
+
+
+def frame_mark(x):
+    x = x.copy()
+    x[:64] = -1
+    x[-64:] = -1
+    x[:, :8] = 1
+    x[:, -8:] = 1
+    return x
+
+
+@torch.inference_mode()
+def pytorch2numpy(imgs):
+    results = []
+    for x in imgs:
+        y = x.movedim(0, -1)
+        y = y * 127.5 + 127.5
+        y = y.detach().float().cpu().numpy().clip(0, 255).astype(np.uint8)
+        results.append(y)
+    return results
+
+
+@torch.inference_mode()
+def numpy2pytorch(imgs):
+    h = torch.from_numpy(np.stack(imgs, axis=0)).float() / 127.5 - 1.0
+    h = h.movedim(-1, 1)
+    return h
+
+
+@torch.no_grad()
+def duplicate_prefix_to_suffix(x, count, zero_out=False):
+    if zero_out:
+        return torch.cat([x, torch.zeros_like(x[:count])], dim=0)
+    else:
+        return torch.cat([x, x[:count]], dim=0)
+
+
+def weighted_mse(a, b, weight):
+    return torch.mean(weight.float() * (a.float() - b.float()) ** 2)
+
+
+def clamped_linear_interpolation(x, x_min, y_min, x_max, y_max, sigma=1.0):
+    x = (x - x_min) / (x_max - x_min)
+    x = max(0.0, min(x, 1.0))
+    x = x ** sigma
+    return y_min + x * (y_max - y_min)
+
+
+def expand_to_dims(x, target_dims):
+    return x.view(*x.shape, *([1] * max(0, target_dims - x.dim())))
+
+
+def repeat_to_batch_size(tensor: torch.Tensor, batch_size: int):
+    if tensor is None:
+        return None
+
+    first_dim = tensor.shape[0]
+
+    if first_dim == batch_size:
+        return tensor
+
+    if batch_size % first_dim != 0:
+        raise ValueError(f"Cannot evenly repeat first dim {first_dim} to match batch_size {batch_size}.")
+
+    repeat_times = batch_size // first_dim
+
+    return tensor.repeat(repeat_times, *[1] * (tensor.dim() - 1))
+
+
+def dim5(x):
+    return expand_to_dims(x, 5)
+
+
+def dim4(x):
+    return expand_to_dims(x, 4)
+
+
+def dim3(x):
+    return expand_to_dims(x, 3)
+
+
+def crop_or_pad_yield_mask(x, length):
+    B, F, C = x.shape
+    device = x.device
+    dtype = x.dtype
+
+    if F < length:
+        y = torch.zeros((B, length, C), dtype=dtype, device=device)
+        mask = torch.zeros((B, length), dtype=torch.bool, device=device)
+        y[:, :F, :] = x
+        mask[:, :F] = True
+        return y, mask
+
+    return x[:, :length, :], torch.ones((B, length), dtype=torch.bool, device=device)
+
+
+def extend_dim(x, dim, minimal_length, zero_pad=False):
+    original_length = int(x.shape[dim])
+
+    if original_length >= minimal_length:
+        return x
+
+    if zero_pad:
+        padding_shape = list(x.shape)
+        padding_shape[dim] = minimal_length - original_length
+        padding = torch.zeros(padding_shape, dtype=x.dtype, device=x.device)
+    else:
+        idx = (slice(None),) * dim + (slice(-1, None),) + (slice(None),) * (len(x.shape) - dim - 1)
+        last_element = x[idx]
+        padding = last_element.repeat_interleave(minimal_length - original_length, dim=dim)
+
+    return torch.cat([x, padding], dim=dim)
+
+
+def lazy_positional_encoding(t, repeats=None):
+    if not isinstance(t, list):
+        t = [t]
+
+    from diffusers.models.embeddings import get_timestep_embedding
+
+    te = torch.tensor(t)
+    te = get_timestep_embedding(timesteps=te, embedding_dim=256, flip_sin_to_cos=True, downscale_freq_shift=0.0, scale=1.0)
+
+    if repeats is None:
+        return te
+
+    te = te[:, None, :].expand(-1, repeats, -1)
+
+    return te
+
+
+def state_dict_offset_merge(A, B, C=None):
+    result = {}
+    keys = A.keys()
+
+    for key in keys:
+        A_value = A[key]
+        B_value = B[key].to(A_value)
+
+        if C is None:
+            result[key] = A_value + B_value
+        else:
+            C_value = C[key].to(A_value)
+            result[key] = A_value + B_value - C_value
+
+    return result
+
+
+def state_dict_weighted_merge(state_dicts, weights):
+    if len(state_dicts) != len(weights):
+        raise ValueError("Number of state dictionaries must match number of weights")
+
+    if not state_dicts:
+        return {}
+
+    total_weight = sum(weights)
+
+    if total_weight == 0:
+        raise ValueError("Sum of weights cannot be zero")
+
+    normalized_weights = [w / total_weight for w in weights]
+
+    keys = state_dicts[0].keys()
+    result = {}
+
+    for key in keys:
+        result[key] = state_dicts[0][key] * normalized_weights[0]
+
+        for i in range(1, len(state_dicts)):
+            state_dict_value = state_dicts[i][key].to(result[key])
+            result[key] += state_dict_value * normalized_weights[i]
+
+    return result
+
+
+def group_files_by_folder(all_files):
+    grouped_files = {}
+
+    for file in all_files:
+        folder_name = os.path.basename(os.path.dirname(file))
+        if folder_name not in grouped_files:
+            grouped_files[folder_name] = []
+        grouped_files[folder_name].append(file)
+
+    list_of_lists = list(grouped_files.values())
+    return list_of_lists
+
+
+def generate_timestamp():
+    now = datetime.datetime.now()
+    timestamp = now.strftime('%y%m%d_%H%M%S')
+    milliseconds = f"{int(now.microsecond / 1000):03d}"
+    random_number = random.randint(0, 9999)
+    return f"{timestamp}_{milliseconds}_{random_number}"
+
+
+def write_PIL_image_with_png_info(image, metadata, path):
+    from PIL.PngImagePlugin import PngInfo
+
+    png_info = PngInfo()
+    for key, value in metadata.items():
+        png_info.add_text(key, value)
+
+    image.save(path, "PNG", pnginfo=png_info)
+    return image
+
+
+def torch_safe_save(content, path):
+    torch.save(content, path + '_tmp')
+    os.replace(path + '_tmp', path)
+    return path
+
+
+def move_optimizer_to_device(optimizer, device):
+    for state in optimizer.state.values():
+        for k, v in state.items():
+            if isinstance(v, torch.Tensor):
+                state[k] = v.to(device)