diff --git a/comfy/ldm/helios/model.py b/comfy/ldm/helios/model.py new file mode 100644 index 000000000..911d45831 --- /dev/null +++ b/comfy/ldm/helios/model.py @@ -0,0 +1,811 @@ +import math + +import torch +import torch.nn as nn + +from comfy.ldm.modules.attention import optimized_attention +from comfy.ldm.flux.layers import EmbedND +from comfy.ldm.flux.math import apply_rope1 +from comfy.ldm.wan.model import sinusoidal_embedding_1d +import comfy.ldm.common_dit +import comfy.patcher_extension + + + +def pad_for_3d_conv(x, kernel_size): + b, c, t, h, w = x.shape + pt, ph, pw = kernel_size + pad_t = (pt - (t % pt)) % pt + pad_h = (ph - (h % ph)) % ph + pad_w = (pw - (w % pw)) % pw + return torch.nn.functional.pad(x, (0, pad_w, 0, pad_h, 0, pad_t), mode="replicate") + + +def center_down_sample_3d(x, kernel_size): + return torch.nn.functional.avg_pool3d(x, kernel_size, stride=kernel_size) + + +class OutputNorm(nn.Module): + + def __init__(self, dim, eps=1e-6, operation_settings={}): + super().__init__() + self.scale_shift_table = nn.Parameter(torch.randn( + 1, + 2, + dim, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) / dim**0.5) + self.norm = operation_settings.get("operations").LayerNorm( + dim, + eps, + elementwise_affine=False, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) + + def forward( + self, + hidden_states: torch.Tensor, + temb: torch.Tensor, + original_context_length: int, + ): + temb = temb[:, -original_context_length:, :] + shift, scale = (self.scale_shift_table.unsqueeze(0).to(temb.device) + temb.unsqueeze(2)).chunk(2, dim=2) + shift = shift.squeeze(2).to(hidden_states.device) + scale = scale.squeeze(2).to(hidden_states.device) + hidden_states = hidden_states[:, -original_context_length:, :] + # Use float32 for numerical stability like diffusers + hidden_states = (self.norm(hidden_states.float()) * (1 + scale) + shift).type_as(hidden_states) + return hidden_states + + +class HeliosSelfAttention(nn.Module): + + def __init__( + self, + dim, + num_heads, + qk_norm=True, + eps=1e-6, + is_cross_attention=False, + is_amplify_history=False, + history_scale_mode="per_head", + operation_settings={}, + ): + super().__init__() + self.num_heads = num_heads + self.head_dim = dim // num_heads + self.is_cross_attention = is_cross_attention + self.is_amplify_history = is_amplify_history + + self.to_q = operation_settings.get("operations").Linear( + dim, + dim, + bias=True, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) + self.to_k = operation_settings.get("operations").Linear( + dim, + dim, + bias=True, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) + self.to_v = operation_settings.get("operations").Linear( + dim, + dim, + bias=True, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) + self.to_out = nn.ModuleList([ + operation_settings.get("operations").Linear( + dim, + dim, + bias=True, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ), + nn.Dropout(0.0), + ]) + + if qk_norm: + self.norm_q = operation_settings.get("operations").RMSNorm( + dim, + eps=eps, + elementwise_affine=True, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) + self.norm_k = operation_settings.get("operations").RMSNorm( + dim, + eps=eps, + elementwise_affine=True, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) + else: + self.norm_q = nn.Identity() + self.norm_k = nn.Identity() + + if is_amplify_history: + if history_scale_mode == "scalar": + self.history_key_scale = nn.Parameter(torch.ones( + 1, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + )) + else: + self.history_key_scale = nn.Parameter(torch.ones( + num_heads, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + )) + self.history_scale_mode = history_scale_mode + self.max_scale = 10.0 + + def forward( + self, + x, + context=None, + freqs=None, + original_context_length=None, + transformer_options={}, + ): + if context is None: + context = x + + b, sq, _ = x.shape + sk = context.shape[1] + + q = self.norm_q(self.to_q(x)).view(b, sq, self.num_heads, self.head_dim) + k = self.norm_k(self.to_k(context)).view(b, sk, self.num_heads, self.head_dim) + v = self.to_v(context).view(b, sk, self.num_heads, self.head_dim) + + if freqs is not None: + q = apply_rope1(q, freqs) + k = apply_rope1(k, freqs) + + if q.dtype != v.dtype: + q = q.to(v.dtype) + if k.dtype != v.dtype: + k = k.to(v.dtype) + + if (not self.is_cross_attention and self.is_amplify_history and original_context_length is not None): + history_seq_len = sq - original_context_length + if history_seq_len > 0: + scale_key = 1.0 + torch.sigmoid(self.history_key_scale) * (self.max_scale - 1.0) + if self.history_scale_mode == "per_head": + scale_key = scale_key.view(1, 1, -1, 1) + k = torch.cat([k[:, :history_seq_len] * scale_key, k[:, history_seq_len:]], dim=1) + + y = optimized_attention( + q.view(b, sq, -1), + k.view(b, sk, -1), + v.view(b, sk, -1), + heads=self.num_heads, + transformer_options=transformer_options, + ) + y = self.to_out[0](y) + y = self.to_out[1](y) + return y + + +class HeliosAttentionBlock(nn.Module): + + def __init__( + self, + dim, + ffn_dim, + num_heads, + qk_norm=True, + cross_attn_norm=True, + eps=1e-6, + guidance_cross_attn=False, + is_amplify_history=False, + history_scale_mode="per_head", + operation_settings={}, + ): + super().__init__() + + self.norm1 = operation_settings.get("operations").LayerNorm( + dim, + eps, + elementwise_affine=False, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) + self.attn1 = HeliosSelfAttention( + dim, + num_heads, + qk_norm=qk_norm, + eps=eps, + is_cross_attention=False, + is_amplify_history=is_amplify_history, + history_scale_mode=history_scale_mode, + operation_settings=operation_settings, + ) + + self.cross_attn_norm = bool(cross_attn_norm) + self.norm2 = (operation_settings.get("operations").LayerNorm( + dim, + eps, + elementwise_affine=True, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) if self.cross_attn_norm else nn.Identity()) + self.attn2 = HeliosSelfAttention( + dim, + num_heads, + qk_norm=qk_norm, + eps=eps, + is_cross_attention=True, + operation_settings=operation_settings, + ) + + self.norm3 = operation_settings.get("operations").LayerNorm( + dim, + eps, + elementwise_affine=False, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) + self.ffn = nn.Sequential( + operation_settings.get("operations").Linear( + dim, + ffn_dim, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ), + nn.GELU(approximate="tanh"), + operation_settings.get("operations").Linear( + ffn_dim, + dim, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ), + ) + + self.scale_shift_table = nn.Parameter(torch.randn( + 1, + 6, + dim, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) / dim**0.5) + self.guidance_cross_attn = guidance_cross_attn + + def forward(self, x, context, e, freqs, original_context_length=None, transformer_options={}): + if e.ndim == 4: + shift_msa, scale_msa, gate_msa, c_shift_msa, c_scale_msa, c_gate_msa = ( + self.scale_shift_table.unsqueeze(0).to(e.device) + e.float() + ).chunk(6, dim=2) + shift_msa = shift_msa.squeeze(2) + scale_msa = scale_msa.squeeze(2) + gate_msa = gate_msa.squeeze(2) + c_shift_msa = c_shift_msa.squeeze(2) + c_scale_msa = c_scale_msa.squeeze(2) + c_gate_msa = c_gate_msa.squeeze(2) + else: + shift_msa, scale_msa, gate_msa, c_shift_msa, c_scale_msa, c_gate_msa = ( + self.scale_shift_table.to(e.device) + e.float() + ).chunk(6, dim=1) + + # self-attn + # Use float32 for numerical stability like diffusers + # norm1 has elementwise_affine=False, so we can safely convert to float32 + norm_x = self.norm1(x.float()) + norm_x = (norm_x * (1 + scale_msa) + shift_msa).type_as(x) + y = self.attn1( + norm_x, + freqs=freqs, + original_context_length=original_context_length, + transformer_options=transformer_options, + ) + x = (x.float() + y.float() * gate_msa).type_as(x) + + # cross-attn + if self.guidance_cross_attn and original_context_length is not None: + history_seq_len = x.shape[1] - original_context_length + history_x, x_main = torch.split(x, [history_seq_len, original_context_length], dim=1) + if self.cross_attn_norm: + # norm2 has elementwise_affine=True, manually do FP32LayerNorm behavior + norm_x_main = torch.nn.functional.layer_norm( + x_main.float(), + self.norm2.normalized_shape, + self.norm2.weight.to(x_main.device).float() if self.norm2.weight is not None else None, + self.norm2.bias.to(x_main.device).float() if self.norm2.bias is not None else None, + self.norm2.eps, + ).type_as(x_main) + else: + norm_x_main = x_main + x_main = x_main + self.attn2( + norm_x_main, + context=context, + transformer_options=transformer_options, + ) + x = torch.cat([history_x, x_main], dim=1) + else: + if self.cross_attn_norm: + # norm2 has elementwise_affine=True, manually do FP32LayerNorm behavior + norm_x = torch.nn.functional.layer_norm( + x.float(), + self.norm2.normalized_shape, + self.norm2.weight.to(x.device).float() if self.norm2.weight is not None else None, + self.norm2.bias.to(x.device).float() if self.norm2.bias is not None else None, + self.norm2.eps, + ).type_as(x) + else: + norm_x = x + x = x + self.attn2(norm_x, context=context, transformer_options=transformer_options) + + # ffn + # Use float32 for numerical stability like diffusers + # norm3 has elementwise_affine=False, so we can safely convert to float32 + norm_x = self.norm3(x.float()) + norm_x = (norm_x * (1 + c_scale_msa) + c_shift_msa).type_as(x) + y = self.ffn(norm_x) + x = (x.float() + y.float() * c_gate_msa).type_as(x) + return x + + +class HeliosModel(torch.nn.Module): + + def __init__( + self, + model_type="t2v", + patch_size=(1, 2, 2), + num_attention_heads=40, + attention_head_dim=128, + in_channels=16, + out_channels=16, + text_dim=4096, + freq_dim=256, + ffn_dim=13824, + num_layers=40, + cross_attn_norm=True, + qk_norm=True, + eps=1e-6, + rope_dim=(44, 42, 42), + rope_theta=10000.0, + guidance_cross_attn=True, + zero_history_timestep=True, + has_multi_term_memory_patch=True, + is_amplify_history=False, + history_scale_mode="per_head", + image_model=None, + device=None, + dtype=None, + operations=None, + **kwargs, + ): + del model_type, image_model, kwargs + super().__init__() + self.dtype = dtype + operation_settings = { + "operations": operations, + "device": device, + "dtype": dtype, + } + + dim = num_attention_heads * attention_head_dim + self.patch_size = patch_size + self.out_dim = out_channels or in_channels + self.dim = dim + self.freq_dim = freq_dim + self.zero_history_timestep = zero_history_timestep + + # embeddings + self.patch_embedding = operations.Conv3d( + in_channels, + dim, + kernel_size=patch_size, + stride=patch_size, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) + self.text_embedding = nn.Sequential( + operations.Linear( + text_dim, + dim, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ), + nn.GELU(approximate="tanh"), + operations.Linear( + dim, + dim, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ), + ) + self.time_embedding = nn.Sequential( + operations.Linear( + freq_dim, + dim, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ), + nn.SiLU(), + operations.Linear( + dim, + dim, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ), + ) + self.time_projection = nn.Sequential( + nn.SiLU(), + operations.Linear( + dim, + dim * 6, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ), + ) + + d = dim // num_attention_heads + self.rope_embedder = EmbedND(dim=d, theta=rope_theta, axes_dim=list(rope_dim)) + + # pyramidal embedding + if has_multi_term_memory_patch: + self.patch_short = operations.Conv3d( + in_channels, + dim, + kernel_size=patch_size, + stride=patch_size, + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) + self.patch_mid = operations.Conv3d( + in_channels, + dim, + kernel_size=tuple(2 * p for p in patch_size), + stride=tuple(2 * p for p in patch_size), + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) + self.patch_long = operations.Conv3d( + in_channels, + dim, + kernel_size=tuple(4 * p for p in patch_size), + stride=tuple(4 * p for p in patch_size), + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) + + # blocks + self.blocks = nn.ModuleList([HeliosAttentionBlock( + dim, + ffn_dim, + num_attention_heads, + qk_norm=qk_norm, + cross_attn_norm=cross_attn_norm, + eps=eps, + guidance_cross_attn=guidance_cross_attn, + is_amplify_history=is_amplify_history, + history_scale_mode=history_scale_mode, + operation_settings=operation_settings, + ) for _ in range(num_layers)]) + + # head + self.norm_out = OutputNorm(dim, eps=eps, operation_settings=operation_settings) + self.proj_out = operations.Linear( + dim, + self.out_dim * math.prod(patch_size), + device=operation_settings.get("device"), + dtype=operation_settings.get("dtype"), + ) + + def rope_encode( + self, + t, + h, + w, + t_start=0, + steps_t=None, + steps_h=None, + steps_w=None, + device=None, + dtype=None, + transformer_options={}, + frame_indices=None, + ): + patch_size = self.patch_size + t_len = (t + (patch_size[0] // 2)) // patch_size[0] + h_len = (h + (patch_size[1] // 2)) // patch_size[1] + w_len = (w + (patch_size[2] // 2)) // patch_size[2] + + if steps_t is None: + steps_t = t_len + if steps_h is None: + steps_h = h_len + if steps_w is None: + steps_w = w_len + + h_start = 0 + w_start = 0 + rope_options = transformer_options.get("rope_options", None) + if rope_options is not None: + t_len = (t_len - 1.0) * rope_options.get("scale_t", 1.0) + 1.0 + h_len = (h_len - 1.0) * rope_options.get("scale_y", 1.0) + 1.0 + w_len = (w_len - 1.0) * rope_options.get("scale_x", 1.0) + 1.0 + + t_start += rope_options.get("shift_t", 0.0) + h_start += rope_options.get("shift_y", 0.0) + w_start += rope_options.get("shift_x", 0.0) + + if frame_indices is None: + t_coords = torch.linspace( + t_start, + t_start + (t_len - 1), + steps=steps_t, + device=device, + dtype=dtype, + ).reshape(1, -1, 1, 1) + batch_size = 1 + else: + batch_size = frame_indices.shape[0] + t_coords = frame_indices.to(device=device, dtype=dtype) + if t_coords.shape[1] != steps_t: + t_coords = torch.nn.functional.interpolate( + t_coords.unsqueeze(1), + size=steps_t, + mode="linear", + align_corners=False, + ).squeeze(1) + t_coords = (t_coords + t_start)[:, :, None, None] + + img_ids = torch.zeros((batch_size, steps_t, steps_h, steps_w, 3), device=device, dtype=dtype) + img_ids[:, :, :, :, 0] = img_ids[:, :, :, :, 0] + t_coords.expand(batch_size, steps_t, steps_h, steps_w) + img_ids[:, :, :, :, 1] = img_ids[:, :, :, :, 1] + torch.linspace(h_start, h_start + (h_len - 1), steps=steps_h, device=device, dtype=dtype).reshape(1, 1, -1, 1) + img_ids[:, :, :, :, 2] = img_ids[:, :, :, :, 2] + torch.linspace(w_start, w_start + (w_len - 1), steps=steps_w, device=device, dtype=dtype).reshape(1, 1, 1, -1) + img_ids = img_ids.reshape(batch_size, -1, img_ids.shape[-1]) + return self.rope_embedder(img_ids).movedim(1, 2) + + def forward( + self, + x, + timestep, + context, + clip_fea=None, + time_dim_concat=None, + transformer_options={}, + **kwargs, + ): + return comfy.patcher_extension.WrapperExecutor.new_class_executor( + self._forward, + self, + comfy.patcher_extension.get_all_wrappers(comfy.patcher_extension.WrappersMP.DIFFUSION_MODEL, transformer_options), + ).execute( + x, + timestep, + context, + clip_fea, + time_dim_concat, + transformer_options, + **kwargs, + ) + + def _forward( + self, + x, + timestep, + context, + clip_fea=None, + time_dim_concat=None, + transformer_options={}, + **kwargs, + ): + del clip_fea, time_dim_concat + + _, _, t, h, w = x.shape + x = comfy.ldm.common_dit.pad_to_patch_size(x, self.patch_size) + + out = self.forward_orig( + hidden_states=x, + timestep=timestep, + context=context, + indices_hidden_states=kwargs.get("indices_hidden_states", None), + indices_latents_history_short=kwargs.get("indices_latents_history_short", None), + indices_latents_history_mid=kwargs.get("indices_latents_history_mid", None), + indices_latents_history_long=kwargs.get("indices_latents_history_long", None), + latents_history_short=kwargs.get("latents_history_short", None), + latents_history_mid=kwargs.get("latents_history_mid", None), + latents_history_long=kwargs.get("latents_history_long", None), + transformer_options=transformer_options, + ) + + return out[:, :, :t, :h, :w] + + def forward_orig( + self, + hidden_states, + timestep, + context, + indices_hidden_states=None, + indices_latents_history_short=None, + indices_latents_history_mid=None, + indices_latents_history_long=None, + latents_history_short=None, + latents_history_mid=None, + latents_history_long=None, + transformer_options={}, + ): + batch_size = hidden_states.shape[0] + p_t, p_h, p_w = self.patch_size + + # embeddings + hidden_states = self.patch_embedding(hidden_states) + _, _, post_t, post_h, post_w = hidden_states.shape + hidden_states = hidden_states.flatten(2).transpose(1, 2) + + if indices_hidden_states is None: + indices_hidden_states = (torch.arange(0, post_t, device=hidden_states.device).unsqueeze(0).expand(batch_size, -1)) + + freqs = self.rope_encode( + t=post_t * self.patch_size[0], + h=post_h * self.patch_size[1], + w=post_w * self.patch_size[2], + steps_t=post_t, + steps_h=post_h, + steps_w=post_w, + device=hidden_states.device, + dtype=hidden_states.dtype, + transformer_options=transformer_options, + frame_indices=indices_hidden_states, + ) + original_context_length = hidden_states.shape[1] + + if latents_history_short is not None and indices_latents_history_short is not None: + x_short = self.patch_short(latents_history_short) + _, _, ts, hs, ws = x_short.shape + x_short = x_short.flatten(2).transpose(1, 2) + f_short = self.rope_encode( + t=ts * self.patch_size[0], + h=hs * self.patch_size[1], + w=ws * self.patch_size[2], + steps_t=ts, + steps_h=hs, + steps_w=ws, + device=x_short.device, + dtype=x_short.dtype, + transformer_options=transformer_options, + frame_indices=indices_latents_history_short, + ) + hidden_states = torch.cat([x_short, hidden_states], dim=1) + freqs = torch.cat([f_short, freqs], dim=1) + + if latents_history_mid is not None and indices_latents_history_mid is not None: + x_mid = self.patch_mid(pad_for_3d_conv(latents_history_mid, (2, 4, 4))) + _, _, tm, hm, wm = x_mid.shape + x_mid = x_mid.flatten(2).transpose(1, 2) + mid_t = indices_latents_history_mid.shape[1] + # patch_mid downsamples by 2 in (t, h, w); build RoPE on the pre-downsample grid. + mid_h = hm * 2 + mid_w = wm * 2 + f_mid = self.rope_encode( + t=mid_t * self.patch_size[0], + h=mid_h * self.patch_size[1], + w=mid_w * self.patch_size[2], + steps_t=mid_t, + steps_h=mid_h, + steps_w=mid_w, + device=x_mid.device, + dtype=x_mid.dtype, + transformer_options=transformer_options, + frame_indices=indices_latents_history_mid, + ) + f_mid = self._rope_downsample_3d(f_mid, (mid_t, mid_h, mid_w), (2, 2, 2)) + hidden_states = torch.cat([x_mid, hidden_states], dim=1) + freqs = torch.cat([f_mid, freqs], dim=1) + + if latents_history_long is not None and indices_latents_history_long is not None: + x_long = self.patch_long(pad_for_3d_conv(latents_history_long, (4, 8, 8))) + _, _, tl, hl, wl = x_long.shape + x_long = x_long.flatten(2).transpose(1, 2) + long_t = indices_latents_history_long.shape[1] + # patch_long downsamples by 4 in (t, h, w); build RoPE on the pre-downsample grid. + long_h = hl * 4 + long_w = wl * 4 + f_long = self.rope_encode( + t=long_t * self.patch_size[0], + h=long_h * self.patch_size[1], + w=long_w * self.patch_size[2], + steps_t=long_t, + steps_h=long_h, + steps_w=long_w, + device=x_long.device, + dtype=x_long.dtype, + transformer_options=transformer_options, + frame_indices=indices_latents_history_long, + ) + f_long = self._rope_downsample_3d(f_long, (long_t, long_h, long_w), (4, 4, 4)) + hidden_states = torch.cat([x_long, hidden_states], dim=1) + freqs = torch.cat([f_long, freqs], dim=1) + + history_context_length = hidden_states.shape[1] - original_context_length + + if timestep.ndim == 0: + timestep = timestep.unsqueeze(0) + timestep = timestep.to(hidden_states.device) + if timestep.shape[0] != batch_size: + timestep = timestep[:1].expand(batch_size) + + # time embeddings + e = self.time_embedding(sinusoidal_embedding_1d(self.freq_dim, timestep.flatten()).to(dtype=hidden_states.dtype)) + e = e.reshape(batch_size, -1, e.shape[-1]) + e0 = self.time_projection(e).unflatten(2, (6, self.dim)) + context = self.text_embedding(context.to(dtype=hidden_states.dtype)) + + if self.zero_history_timestep and history_context_length > 0: + timestep_t0 = torch.zeros((1, ), dtype=timestep.dtype, device=timestep.device) + e_t0 = self.time_embedding(sinusoidal_embedding_1d(self.freq_dim, timestep_t0.flatten()).to(dtype=hidden_states.dtype)) + e_t0 = e_t0.reshape(1, -1, e_t0.shape[-1]).expand(batch_size, history_context_length, -1) + e0_t0 = self.time_projection(e_t0[:, :1]).unflatten(2, (6, self.dim)) + e0_t0 = (e0_t0.view(batch_size, 1, 6, self.dim).permute(0, 2, 1, 3).expand(batch_size, 6, history_context_length, self.dim)) + + e = e.expand(batch_size, original_context_length, -1) + e0 = (e0.view(batch_size, 1, 6, self.dim).permute(0, 2, 1, 3).expand(batch_size, 6, original_context_length, self.dim)) + e = torch.cat([e_t0, e], dim=1) + e0 = torch.cat([e0_t0, e0], dim=2) + else: + e = e.expand(batch_size, hidden_states.shape[1], -1) + e0 = (e0.view(batch_size, 1, 6, self.dim).permute(0, 2, 1, 3).expand(batch_size, 6, hidden_states.shape[1], self.dim)) + + e0 = e0.permute(0, 2, 1, 3) + + for i_b, block in enumerate(self.blocks): + hidden_states = block( + hidden_states, + context, + e0, + freqs, + original_context_length=original_context_length, + transformer_options=transformer_options, + ) + hidden_states = self.norm_out(hidden_states, e, original_context_length) + hidden_states = self.proj_out(hidden_states) + return self.unpatchify(hidden_states, (post_t, post_h, post_w)) + + def unpatchify(self, x, grid_sizes): + """ + Unpatchify the output from proj_out back to video format. + + Args: + x: [batch, num_patches, out_dim * prod(patch_size)] + grid_sizes: (num_frames, height, width) in patch space + + Returns: + [batch, out_dim, num_frames, height, width] in pixel space + """ + b = x.shape[0] + post_t, post_h, post_w = grid_sizes + p_t, p_h, p_w = self.patch_size + + # Reshape: [B, T*H*W, out_dim*p_t*p_h*p_w] -> [B, T, H, W, p_t, p_h, p_w, out_dim] + # Use -1 to let PyTorch infer the channel dimension (out_dim) + hidden_states = x.reshape(b, post_t, post_h, post_w, p_t, p_h, p_w, -1) + + # Permute: [B, T, H, W, p_t, p_h, p_w, C] -> [B, C, T, p_t, H, p_h, W, p_w] + hidden_states = hidden_states.permute(0, 7, 1, 4, 2, 5, 3, 6) + + # Flatten patches: [B, C, T, p_t, H, p_h, W, p_w] -> [B, C, T*p_t, H*p_h, W*p_w] + output = hidden_states.flatten(6, 7).flatten(4, 5).flatten(2, 3) + + return output + def _rope_downsample_3d(self, freqs, grid_sizes, kernel_size): + b, _, one, d, i2, j2 = freqs.shape + gt, gh, gw = grid_sizes + c = one * d * i2 * j2 + freqs_3d = freqs.reshape(b, gt, gh, gw, c).permute(0, 4, 1, 2, 3) + freqs_3d = pad_for_3d_conv(freqs_3d, kernel_size) + freqs_3d = center_down_sample_3d(freqs_3d, kernel_size) + dt, dh, dw = freqs_3d.shape[2:] + freqs_3d = freqs_3d.permute(0, 2, 3, 4, 1).reshape(b, dt * dh * dw, one, d, i2, j2) + return freqs_3d + +# Backward-compatible alias for existing integration points. +HeliosTransformer3DModel = HeliosModel diff --git a/comfy/model_base.py b/comfy/model_base.py index d9d5a9293..bd0541c02 100644 --- a/comfy/model_base.py +++ b/comfy/model_base.py @@ -41,6 +41,7 @@ import comfy.ldm.cosmos.predict2 import comfy.ldm.lumina.model import comfy.ldm.wan.model import comfy.ldm.wan.model_animate +import comfy.ldm.helios.model import comfy.ldm.hunyuan3d.model import comfy.ldm.hidream.model import comfy.ldm.chroma.model @@ -1268,6 +1269,70 @@ class ZImagePixelSpace(Lumina2): BaseModel.__init__(self, model_config, model_type, device=device, unet_model=comfy.ldm.lumina.model.NextDiTPixelSpace) self.memory_usage_factor_conds = ("ref_latents",) +class Helios(BaseModel): + def __init__(self, model_config, model_type=ModelType.FLOW, device=None): + super().__init__(model_config, model_type, device=device, unet_model=comfy.ldm.helios.model.HeliosTransformer3DModel) + + def extra_conds(self, **kwargs): + out = super().extra_conds(**kwargs) + cross_attn = kwargs.get("cross_attn", None) + if cross_attn is not None: + out["c_crossattn"] = comfy.conds.CONDRegular(cross_attn) + + cond_keys = ( + "indices_hidden_states", + "indices_latents_history_short", + "indices_latents_history_mid", + "indices_latents_history_long", + "latents_history_short", + "latents_history_mid", + "latents_history_long", + "helios_stage_sigmas", + "helios_stage_timesteps", + ) + + for key in cond_keys: + value = kwargs.get(key, None) + if value is None: + continue + # Diffusers forwards Helios history latents without latent-format re-normalization. + # Keep raw history tensors to match transformer inputs across frameworks. + if key in ("helios_stage_sigmas", "helios_stage_timesteps"): + out[key] = comfy.conds.CONDConstant(value) + else: + out[key] = comfy.conds.CONDRegular(value) + return out + + def process_timestep(self, timestep, **kwargs): + stage_sigmas = kwargs.get("helios_stage_sigmas", None) + stage_timesteps = kwargs.get("helios_stage_timesteps", None) + if stage_sigmas is None or stage_timesteps is None: + return timestep + + if stage_sigmas.ndim > 1: + stage_sigmas = stage_sigmas[0] + if stage_timesteps.ndim > 1: + stage_timesteps = stage_timesteps[0] + + if stage_timesteps.numel() == 0 or stage_sigmas.numel() == 0: + return timestep + + if stage_sigmas.numel() == stage_timesteps.numel() + 1: + sigma_candidates = stage_sigmas[:-1] + else: + sigma_candidates = stage_sigmas[: stage_timesteps.numel()] + + if sigma_candidates.numel() == 0: + return timestep + + multiplier = float(getattr(self.model_sampling, "multiplier", 1000.0)) + sigma_in = timestep / multiplier + idx = torch.argmin(torch.abs(sigma_in.unsqueeze(-1) - sigma_candidates.unsqueeze(0)), dim=-1) + mapped = stage_timesteps[idx].to(dtype=timestep.dtype) + if mapped.dtype.is_floating_point: + mapped = torch.floor(mapped) + return mapped + class WAN21(BaseModel): def __init__(self, model_config, model_type=ModelType.FLOW, image_to_video=False, device=None): super().__init__(model_config, model_type, device=device, unet_model=comfy.ldm.wan.model.WanModel) diff --git a/comfy/model_detection.py b/comfy/model_detection.py index 35a6822e3..fffda0db9 100644 --- a/comfy/model_detection.py +++ b/comfy/model_detection.py @@ -490,6 +490,54 @@ def detect_unet_config(state_dict, key_prefix, metadata=None): return dit_config + helios_required_keys = ( + '{}patch_mid.weight'.format(key_prefix), + '{}patch_long.weight'.format(key_prefix), + ) + if all(k in state_dict_keys for k in helios_required_keys): # Helios + dit_config = {} + dit_config["image_model"] = "helios" + + patch_weight = state_dict['{}patch_embedding.weight'.format(key_prefix)] + inner_dim = patch_weight.shape[0] + patch_size = tuple(patch_weight.shape[2:]) + out_proj = state_dict['{}proj_out.weight'.format(key_prefix)] + + dit_config["patch_size"] = patch_size + dit_config["in_channels"] = patch_weight.shape[1] + dit_config["out_channels"] = out_proj.shape[0] // math.prod(patch_size) + text_w = state_dict['{}text_embedding.0.weight'.format(key_prefix)] + time_w = state_dict['{}time_embedding.0.weight'.format(key_prefix)] + dit_config["text_dim"] = text_w.shape[1] + dit_config["freq_dim"] = time_w.shape[1] + dit_config["num_layers"] = count_blocks(state_dict_keys, '{}blocks.'.format(key_prefix) + '{}.') + dit_config["num_attention_heads"] = inner_dim // 128 + dit_config["attention_head_dim"] = 128 + + ffn_in = state_dict.get('{}blocks.0.ffn.net.0.proj.weight'.format(key_prefix), None) + if ffn_in is None: + ffn_in = state_dict.get('{}blocks.0.ffn.0.weight'.format(key_prefix), None) + if ffn_in is not None: + dit_config["ffn_dim"] = ffn_in.shape[0] + + if '{}blocks.0.attn2.add_k_proj.weight'.format(key_prefix) in state_dict_keys: + dit_config["added_kv_proj_dim"] = state_dict['{}blocks.0.attn2.add_k_proj.weight'.format(key_prefix)].shape[1] + + if '{}patch_short.weight'.format(key_prefix) in state_dict_keys: + dit_config["has_multi_term_memory_patch"] = True + else: + dit_config["has_multi_term_memory_patch"] = False + + if '{}blocks.0.attn1.history_key_scale'.format(key_prefix) in state_dict_keys: + dit_config["is_amplify_history"] = True + hk = state_dict['{}blocks.0.attn1.history_key_scale'.format(key_prefix)] + dit_config["history_scale_mode"] = "per_head" if len(hk.shape) > 0 and hk.numel() > 1 else "scalar" + + if metadata is not None and "config" in metadata: + dit_config.update(json.loads(metadata["config"]).get("transformer", {})) + + return dit_config + if '{}head.modulation'.format(key_prefix) in state_dict_keys: # Wan 2.1 dit_config = {} dit_config["image_model"] = "wan2.1" diff --git a/comfy/sd.py b/comfy/sd.py index 4d427bb9a..afd6e6d27 100644 --- a/comfy/sd.py +++ b/comfy/sd.py @@ -1168,6 +1168,7 @@ class CLIPType(Enum): NEWBIE = 24 FLUX2 = 25 LONGCAT_IMAGE = 26 + HELIOS = 27 @@ -1339,6 +1340,11 @@ def load_text_encoder_state_dicts(state_dicts=[], embedding_directory=None, clip clip_target.clip = comfy.text_encoders.wan.te(**t5xxl_detect(clip_data)) clip_target.tokenizer = comfy.text_encoders.wan.WanT5Tokenizer tokenizer_data["spiece_model"] = clip_data[0].get("spiece_model", None) + elif clip_type == CLIPType.HELIOS: + # Helios reuses the WAN UMT5-XXL text encoder stack. + clip_target.clip = comfy.text_encoders.wan.te(**t5xxl_detect(clip_data)) + clip_target.tokenizer = comfy.text_encoders.wan.WanT5Tokenizer + tokenizer_data["spiece_model"] = clip_data[0].get("spiece_model", None) elif clip_type == CLIPType.HIDREAM: clip_target.clip = comfy.text_encoders.hidream.hidream_clip(**t5xxl_detect(clip_data), clip_l=False, clip_g=False, t5=True, llama=False, dtype_llama=None) diff --git a/comfy/supported_models.py b/comfy/supported_models.py index 07feb31b3..5f58e0a9f 100644 --- a/comfy/supported_models.py +++ b/comfy/supported_models.py @@ -1132,6 +1132,42 @@ class ZImagePixelSpace(ZImage): def get_model(self, state_dict, prefix="", device=None): return model_base.ZImagePixelSpace(self, device=device) +class Helios(supported_models_base.BASE): + unet_config = { + "image_model": "helios", + } + + sampling_settings = { + "shift": 1.0, + } + + unet_extra_config = {} + latent_format = latent_formats.Wan21 + memory_usage_factor = 1.8 + supported_inference_dtypes = [torch.bfloat16, torch.float16, torch.float32] + + vae_key_prefix = ["vae."] + text_encoder_key_prefix = ["text_encoders."] + + def __init__(self, unet_config): + super().__init__(unet_config) + self.memory_usage_factor = (self.unet_config.get("num_layers", 40) * self.unet_config.get("num_attention_heads", 40)) / (40 * 40) * 1.8 + + def get_model(self, state_dict, prefix="", device=None): + return model_base.Helios(self, device=device) + + def clip_target(self, state_dict={}): + pref = self.text_encoder_key_prefix[0] + t5_detect = comfy.text_encoders.sd3_clip.t5_xxl_detect( + state_dict, + "{}umt5xxl.transformer.".format(pref), + ) + # Directly reuse WAN text encoder stack; no Helios-specific TE. + return supported_models_base.ClipTarget( + comfy.text_encoders.wan.WanT5Tokenizer, + comfy.text_encoders.wan.te(**t5_detect), + ) + class WAN21_T2V(supported_models_base.BASE): unet_config = { "image_model": "wan2.1", @@ -1734,6 +1770,6 @@ class LongCatImage(supported_models_base.BASE): hunyuan_detect = comfy.text_encoders.hunyuan_video.llama_detect(state_dict, "{}qwen25_7b.transformer.".format(pref)) return supported_models_base.ClipTarget(comfy.text_encoders.longcat_image.LongCatImageTokenizer, comfy.text_encoders.longcat_image.te(**hunyuan_detect)) -models = [LotusD, Stable_Zero123, SD15_instructpix2pix, SD15, SD20, SD21UnclipL, SD21UnclipH, SDXL_instructpix2pix, SDXLRefiner, SDXL, SSD1B, KOALA_700M, KOALA_1B, Segmind_Vega, SD_X4Upscaler, Stable_Cascade_C, Stable_Cascade_B, SV3D_u, SV3D_p, SD3, StableAudio, AuraFlow, PixArtAlpha, PixArtSigma, HunyuanDiT, HunyuanDiT1, FluxInpaint, Flux, LongCatImage, FluxSchnell, GenmoMochi, LTXV, LTXAV, HunyuanVideo15_SR_Distilled, HunyuanVideo15, HunyuanImage21Refiner, HunyuanImage21, HunyuanVideoSkyreelsI2V, HunyuanVideoI2V, HunyuanVideo, CosmosT2V, CosmosI2V, CosmosT2IPredict2, CosmosI2VPredict2, ZImagePixelSpace, ZImage, Lumina2, WAN22_T2V, WAN21_T2V, WAN21_I2V, WAN21_FunControl2V, WAN21_Vace, WAN21_Camera, WAN22_Camera, WAN22_S2V, WAN21_HuMo, WAN22_Animate, WAN21_FlowRVS, WAN21_SCAIL, Hunyuan3Dv2mini, Hunyuan3Dv2, Hunyuan3Dv2_1, HiDream, Chroma, ChromaRadiance, ACEStep, ACEStep15, Omnigen2, QwenImage, Flux2, Kandinsky5Image, Kandinsky5, Anima] +models = [LotusD, Stable_Zero123, SD15_instructpix2pix, SD15, SD20, SD21UnclipL, SD21UnclipH, SDXL_instructpix2pix, SDXLRefiner, SDXL, SSD1B, KOALA_700M, KOALA_1B, Segmind_Vega, SD_X4Upscaler, Stable_Cascade_C, Stable_Cascade_B, SV3D_u, SV3D_p, SD3, StableAudio, AuraFlow, PixArtAlpha, PixArtSigma, HunyuanDiT, HunyuanDiT1, FluxInpaint, Flux, LongCatImage, FluxSchnell, GenmoMochi, LTXV, LTXAV, HunyuanVideo15_SR_Distilled, HunyuanVideo15, HunyuanImage21Refiner, HunyuanImage21, HunyuanVideoSkyreelsI2V, HunyuanVideoI2V, HunyuanVideo, CosmosT2V, CosmosI2V, CosmosT2IPredict2, CosmosI2VPredict2, ZImagePixelSpace, ZImage, Lumina2, Helios, WAN22_T2V, WAN21_T2V, WAN21_I2V, WAN21_FunControl2V, WAN21_Vace, WAN21_Camera, WAN22_Camera, WAN22_S2V, WAN21_HuMo, WAN22_Animate, WAN21_FlowRVS, WAN21_SCAIL, Hunyuan3Dv2mini, Hunyuan3Dv2, Hunyuan3Dv2_1, HiDream, Chroma, ChromaRadiance, ACEStep, ACEStep15, Omnigen2, QwenImage, Flux2, Kandinsky5Image, Kandinsky5, Anima] models += [SVD_img2vid] diff --git a/comfy_extras/nodes_helios.py b/comfy_extras/nodes_helios.py new file mode 100644 index 000000000..2a356be68 --- /dev/null +++ b/comfy_extras/nodes_helios.py @@ -0,0 +1,1291 @@ +import math +import torch + +import nodes +import comfy.model_management +import comfy.model_patcher +import comfy.sample +import comfy.samplers +import comfy.utils +import comfy.latent_formats +import latent_preview +import node_helpers + +from typing_extensions import override +from comfy_api.latest import ComfyExtension, io + + + + + +def _parse_int_list(values, default): + if values is None: + return default + if isinstance(values, (list, tuple)): + out = [] + for v in values: + try: + out.append(int(v)) + except Exception: + pass + return out if len(out) > 0 else default + + parts = [x.strip() for x in str(values).replace(";", ",").split(",")] + out = [] + for p in parts: + if len(p) == 0: + continue + try: + out.append(int(p)) + except Exception: + continue + return out if len(out) > 0 else default + + +_HELIOS_LATENT_FORMAT = comfy.latent_formats.Wan21() + + +def _apply_helios_latent_space_noise(latent, sigma, generator=None): + """Apply noise in Helios model latent space, then map back to VAE latent space.""" + latent_in = _HELIOS_LATENT_FORMAT.process_in(latent) + noise = torch.randn( + latent_in.shape, + device=latent_in.device, + dtype=latent_in.dtype, + generator=generator, + ) + noised_in = sigma * noise + (1.0 - sigma) * latent_in + return _HELIOS_LATENT_FORMAT.process_out(noised_in).to(device=latent.device, dtype=latent.dtype) + + +def _parse_float_list(values, default): + if values is None: + return default + if isinstance(values, (list, tuple)): + out = [] + for v in values: + try: + out.append(float(v)) + except Exception: + pass + return out if len(out) > 0 else default + + parts = [x.strip() for x in str(values).replace(";", ",").split(",")] + out = [] + for p in parts: + if len(p) == 0: + continue + try: + out.append(float(p)) + except Exception: + continue + return out if len(out) > 0 else default + + +def _strict_bool(value, default=False): + if isinstance(value, bool): + return value + if isinstance(value, int): + return value != 0 + # Reject non-bool numerics from stale workflows (e.g. 0.135). + return bool(default) + + +def _extract_condition_value(conditioning, key): + for c in conditioning: + if len(c) < 2: + continue + value = c[1].get(key, None) + if value is not None: + return value + return None + + +def _process_latent_in_preserve_zero_frames(model, latent, valid_mask=None): + if latent is None or len(latent.shape) != 5: + return latent + if valid_mask is None: + raise ValueError("Helios requires `helios_history_valid_mask` for history latent conversion.") + vm = valid_mask + if not torch.is_tensor(vm): + vm = torch.tensor(vm, device=latent.device) + vm = vm.to(device=latent.device) + if vm.ndim == 2: + nonzero = vm.any(dim=0) + else: + nonzero = vm.reshape(-1) + nonzero = nonzero.bool() + + if nonzero.numel() == 0 or (not torch.any(nonzero)): + return latent + + if nonzero.shape[0] != latent.shape[2]: + raise ValueError( + f"Helios history mask length mismatch: mask_t={nonzero.shape[0]} latent_t={latent.shape[2]}" + ) + + converted = model.model.process_latent_in(latent) + out = latent.clone() + out[:, :, nonzero, :, :] = converted[:, :, nonzero, :, :] + return out + + +def _upsample_latent_5d(latent, scale=2): + b, c, t, h, w = latent.shape + x = latent.permute(0, 2, 1, 3, 4).reshape(b * t, c, h, w) + x = comfy.utils.common_upscale(x, w * scale, h * scale, "nearest", "disabled") + x = x.reshape(b, t, c, h * scale, w * scale).permute(0, 2, 1, 3, 4) + return x + + +def _downsample_latent_5d_bilinear_x2(latent): + b, c, t, h, w = latent.shape + x = latent.permute(0, 2, 1, 3, 4).reshape(b * t, c, h, w) + x = comfy.utils.common_upscale(x, max(1, w // 2), max(1, h // 2), "bilinear", "disabled") * 2.0 + x = x.reshape(b, t, c, max(1, h // 2), max(1, w // 2)).permute(0, 2, 1, 3, 4) + return x + + +def _prepare_stage0_latent(batch, channels, frames, height, width, stage_count, add_noise, seed, dtype, layout, device): + """Prepare initial latent for stage 0 with optional noise""" + full_latent = torch.zeros((batch, channels, frames, height, width), dtype=dtype, layout=layout, device=device) + if add_noise: + full_latent = comfy.sample.prepare_noise(full_latent, seed).to(dtype) + + # Downsample to stage 0 resolution + stage_latent = full_latent + for _ in range(max(0, int(stage_count) - 1)): + stage_latent = _downsample_latent_5d_bilinear_x2(stage_latent) + return stage_latent + + +def _downsample_latent_for_stage0(latent, stage_count): + """Downsample latent to stage 0 resolution.""" + stage_latent = latent + for _ in range(max(0, int(stage_count) - 1)): + stage_latent = _downsample_latent_5d_bilinear_x2(stage_latent) + return stage_latent + + + +def _sample_block_noise_like(latent, gamma, patch_size=(1, 2, 2), generator=None, seed=None): + b, c, t, h, w = latent.shape + _, ph, pw = patch_size + block_size = ph * pw + + cov = torch.eye(block_size, device=latent.device) * (1.0 + gamma) - torch.ones(block_size, block_size, device=latent.device) * gamma + cov += torch.eye(block_size, device=latent.device) * 1e-6 + + h_blocks = h // ph + w_blocks = w // pw + block_number = b * c * t * h_blocks * w_blocks + + if generator is not None: + # Exact sampling path (MultivariateNormal.sample), while consuming + # from an explicit generator by temporarily swapping default RNG state. + with torch.random.fork_rng(devices=[latent.device] if latent.device.type == "cuda" else []): + if latent.device.type == "cuda": + torch.cuda.set_rng_state(generator.get_state(), device=latent.device) + else: + torch.random.set_rng_state(generator.get_state()) + dist = torch.distributions.MultivariateNormal( + torch.zeros(block_size, device=latent.device), + covariance_matrix=cov, + ) + noise = dist.sample((block_number,)) + if latent.device.type == "cuda": + generator.set_state(torch.cuda.get_rng_state(device=latent.device)) + else: + generator.set_state(torch.random.get_rng_state()) + elif seed is None: + dist = torch.distributions.MultivariateNormal(torch.zeros(block_size, device=latent.device), covariance_matrix=cov) + noise = dist.sample((block_number,)) + else: + # Use deterministic RNG when seed is provided (for cross-framework alignment). + with torch.random.fork_rng(devices=[latent.device] if latent.device.type == "cuda" else []): + torch.manual_seed(int(seed)) + dist = torch.distributions.MultivariateNormal(torch.zeros(block_size, device=latent.device), covariance_matrix=cov) + noise = dist.sample((block_number,)) + noise = noise.view(b, c, t, h_blocks, w_blocks, ph, pw) + noise = noise.permute(0, 1, 2, 3, 5, 4, 6).reshape(b, c, t, h, w) + return noise + + +def _helios_global_sigmas(num_train_timesteps=1000, shift=1.0): + alphas = torch.linspace(1.0, 1.0 / float(num_train_timesteps), num_train_timesteps + 1) + sigmas = 1.0 - alphas + if abs(shift - 1.0) > 1e-8: + sigmas = shift * sigmas / (1.0 + (shift - 1.0) * sigmas) + return torch.flip(sigmas, dims=[0])[:-1] + + +def _helios_stage_tables(stage_count, stage_range, gamma, num_train_timesteps=1000, shift=1.0): + sigmas = _helios_global_sigmas(num_train_timesteps=num_train_timesteps, shift=shift) + + ori_start_sigmas = {} + start_sigmas = {} + end_sigmas = {} + timestep_ratios = {} + timesteps_per_stage = {} + sigmas_per_stage = {} + + stage_distance = [] + for i in range(stage_count): + start_indice = int(max(0.0, min(1.0, stage_range[i])) * num_train_timesteps) + end_indice = int(max(0.0, min(1.0, stage_range[i + 1])) * num_train_timesteps) + start_indice = max(0, min(num_train_timesteps - 1, start_indice)) + end_indice = max(0, min(num_train_timesteps, end_indice)) + + start_sigma = float(sigmas[start_indice].item()) + end_sigma = float(sigmas[end_indice].item()) if end_indice < num_train_timesteps else 0.0 + ori_start_sigmas[i] = start_sigma + + if i != 0: + ori_sigma = 1.0 - start_sigma + corrected_sigma = (1.0 / (math.sqrt(1.0 + (1.0 / gamma)) * (1.0 - ori_sigma) + ori_sigma)) * ori_sigma + start_sigma = 1.0 - corrected_sigma + + stage_distance.append(start_sigma - end_sigma) + start_sigmas[i] = start_sigma + end_sigmas[i] = end_sigma + + tot_distance = sum(stage_distance) if sum(stage_distance) > 1e-12 else 1.0 + for i in range(stage_count): + start_ratio = 0.0 if i == 0 else sum(stage_distance[:i]) / tot_distance + end_ratio = 0.9999999999999999 if i == stage_count - 1 else sum(stage_distance[: i + 1]) / tot_distance + timestep_ratios[i] = (start_ratio, end_ratio) + + tmax = min(float(sigmas[int(start_ratio * num_train_timesteps)].item() * num_train_timesteps), 999.0) + tmin = float(sigmas[min(int(end_ratio * num_train_timesteps), num_train_timesteps - 1)].item() * num_train_timesteps) + timesteps_per_stage[i] = torch.linspace(tmax, tmin, num_train_timesteps + 1)[:-1] + # Fixed: use the same sigma range [0.999, 0] for all stages. + sigmas_per_stage[i] = torch.linspace(0.999, 0.0, num_train_timesteps + 1)[:-1] + + + return { + "ori_start_sigmas": ori_start_sigmas, + "start_sigmas": start_sigmas, + "end_sigmas": end_sigmas, + "timestep_ratios": timestep_ratios, + "timesteps_per_stage": timesteps_per_stage, + "sigmas_per_stage": sigmas_per_stage, + } + + +def _helios_stage_sigmas(stage_idx, stage_steps, stage_tables, is_distilled=False, is_amplify_first_stage=False): + stage_steps = max(1, int(stage_steps)) + if is_distilled: + stage_steps = stage_steps * 2 if (is_amplify_first_stage and stage_idx == 0) else stage_steps + + stage_sigma_src = stage_tables["sigmas_per_stage"][stage_idx] + sigmas = torch.linspace(float(stage_sigma_src[0].item()), float(stage_sigma_src[-1].item()), stage_steps) + sigmas = torch.cat([sigmas, torch.zeros(1, dtype=sigmas.dtype, device=sigmas.device)], dim=0) + return sigmas + + +def _helios_stage_timesteps(stage_idx, stage_steps, stage_tables, is_distilled=False, is_amplify_first_stage=False): + stage_steps = max(1, int(stage_steps)) + if is_distilled: + stage_steps = stage_steps * 2 if (is_amplify_first_stage and stage_idx == 0) else stage_steps + + stage_timestep_src = stage_tables["timesteps_per_stage"][stage_idx] + timesteps = torch.linspace(float(stage_timestep_src[0].item()), float(stage_timestep_src[-1].item()), stage_steps) + return timesteps + + +def _calculate_shift(image_seq_len, base_seq_len=256, max_seq_len=4096, base_shift=0.5, max_shift=1.15): + m = (max_shift - base_shift) / float(max_seq_len - base_seq_len) + b = base_shift - m * float(base_seq_len) + return float(image_seq_len) * m + b + + +def _time_shift_linear(mu, sigma, t): + return mu / (mu + (1.0 / t - 1.0) ** sigma) + + +def _time_shift_exponential(mu, sigma, t): + return math.exp(mu) / (math.exp(mu) + (1.0 / t - 1.0) ** sigma) + + +def _time_shift(t, mu, sigma=1.0, mode="exponential"): + t = torch.clamp(t, min=1e-6, max=0.999999) + if mode == "linear": + return _time_shift_linear(mu, sigma, t) + return _time_shift_exponential(mu, sigma, t) + + +def _optimized_scale(positive_flat, negative_flat): + dot_product = torch.sum(positive_flat * negative_flat, dim=1, keepdim=True) + squared_norm = torch.sum(negative_flat * negative_flat, dim=1, keepdim=True) + 1e-8 + return dot_product / squared_norm + + +def _build_cfg_zero_star_pre_cfg(stage_idx, zero_steps, use_zero_init): + state = {"i": 0} + + def pre_cfg_fn(args): + conds_out = args["conds_out"] + if len(conds_out) < 2 or conds_out[1] is None: + state["i"] += 1 + return conds_out + + denoised_text = conds_out[0] + denoised_uncond = conds_out[1] + cfg = float(args.get("cond_scale", 1.0)) + x = args["input"] + sigma = args["sigma"] + + sigma_reshaped = sigma.reshape(sigma.shape[0], *([1] * (denoised_text.ndim - 1))) + sigma_safe = torch.clamp(sigma_reshaped, min=1e-8) + + flow_text = (x - denoised_text) / sigma_safe + flow_uncond = (x - denoised_uncond) / sigma_safe + + positive_flat = flow_text.reshape(flow_text.shape[0], -1) + negative_flat = flow_uncond.reshape(flow_uncond.shape[0], -1) + alpha = _optimized_scale(positive_flat, negative_flat) + alpha = alpha.reshape(flow_text.shape[0], *([1] * (flow_text.ndim - 1))).to(flow_text.dtype) + + if stage_idx == 0 and state["i"] <= int(zero_steps) and bool(use_zero_init): + flow_final = flow_text * 0.0 + else: + flow_final = flow_uncond * alpha + cfg * (flow_text - flow_uncond * alpha) + + denoised_final = x - flow_final * sigma_safe + + state["i"] += 1 + return [denoised_final, denoised_final] + + return pre_cfg_fn + + +def _helios_euler_sample(model, x, sigmas, extra_args=None, callback=None, disable=None): + extra_args = {} if extra_args is None else extra_args + s_in = x.new_ones([x.shape[0]]) + + for i in range(len(sigmas) - 1): + sigma = sigmas[i] + sigma_next = sigmas[i + 1] + denoised = model(x, sigma * s_in, **extra_args) + + sigma_safe = sigma if float(sigma) > 1e-8 else sigma.new_tensor(1e-8) + flow_pred = (x - denoised) / sigma_safe + + if callback is not None: + callback({"x": x, "i": i, "sigma": sigma, "sigma_hat": sigma, "denoised": denoised}) + + x = x + (sigma_next - sigma) * flow_pred + + return x + + +def _helios_dmd_sample( + model, + x, + sigmas, + extra_args=None, + callback=None, + disable=None, + dmd_noisy_tensor=None, + dmd_sigmas=None, + dmd_timesteps=None, + all_timesteps=None, +): + extra_args = {} if extra_args is None else extra_args + s_in = x.new_ones([x.shape[0]]) + if dmd_noisy_tensor is None: + dmd_noisy_tensor = x + dmd_noisy_tensor = dmd_noisy_tensor.to(device=x.device, dtype=x.dtype) + if dmd_sigmas is None: + dmd_sigmas = sigmas + if dmd_timesteps is None: + dmd_timesteps = torch.arange(len(sigmas) - 1, device=sigmas.device, dtype=sigmas.dtype) + if all_timesteps is None: + all_timesteps = dmd_timesteps + + def timestep_to_sigma(t): + dt = dmd_timesteps.to(device=x.device, dtype=x.dtype) + ds = dmd_sigmas.to(device=x.device, dtype=x.dtype) + tid = torch.argmin(torch.abs(dt - t)) + tid = torch.clamp(tid, min=0, max=ds.shape[0] - 1) + return ds[tid] + + for i in range(len(sigmas) - 1): + sigma = sigmas[i] + timestep = all_timesteps[i] if i < len(all_timesteps) else i + denoised = model(x, sigma * s_in, **extra_args) + + if callback is not None: + callback({"x": x, "i": i, "sigma": sigma, "sigma_hat": sigma, "denoised": denoised}) + + if i < (len(sigmas) - 2): + timestep_next = all_timesteps[i + 1] if i + 1 < len(all_timesteps) else (i + 1) + sigma_t = timestep_to_sigma(torch.as_tensor(timestep, device=x.device, dtype=x.dtype)) + sigma_next_t = timestep_to_sigma(torch.as_tensor(timestep_next, device=x.device, dtype=x.dtype)) + x0_pred = x - sigma_t * ((x - denoised) / torch.clamp(sigma_t, min=1e-8)) + x = (1.0 - sigma_next_t) * x0_pred + sigma_next_t * dmd_noisy_tensor + else: + x = denoised + + return x + + +def _set_helios_history_values(positive, negative, history_latent, history_sizes, keep_first_frame, prefix_latent=None): + latent = history_latent + if latent is None or len(latent.shape) != 5: + return positive, negative + if prefix_latent is not None and (latent.device != prefix_latent.device or latent.dtype != prefix_latent.dtype): + latent = latent.to(device=prefix_latent.device, dtype=prefix_latent.dtype) + + sizes = list(history_sizes) + if len(sizes) != 3: + sizes = [16, 2, 1] + sizes = [max(0, int(v)) for v in sizes] + total = sum(sizes) + if total <= 0: + return positive, negative + + if latent.shape[2] < total: + pad = torch.zeros( + latent.shape[0], + latent.shape[1], + total - latent.shape[2], + latent.shape[3], + latent.shape[4], + device=latent.device, + dtype=latent.dtype, + ) + hist = torch.cat([pad, latent], dim=2) + else: + hist = latent[:, :, -total:] + + latents_history_long, latents_history_mid, latents_history_short_base = hist.split(sizes, dim=2) + + if keep_first_frame: + if prefix_latent is not None: + prefix = prefix_latent + elif latent.shape[2] > 0: + prefix = latent[:, :, :1] + else: + prefix = torch.zeros(latent.shape[0], latent.shape[1], 1, latent.shape[3], latent.shape[4], device=latent.device, dtype=latent.dtype) + if prefix.device != latents_history_short_base.device or prefix.dtype != latents_history_short_base.dtype: + prefix = prefix.to(device=latents_history_short_base.device, dtype=latents_history_short_base.dtype) + latents_history_short = torch.cat([prefix, latents_history_short_base], dim=2) + else: + latents_history_short = latents_history_short_base + + idx_short = torch.arange(latents_history_short.shape[2], device=latent.device, dtype=torch.int64).unsqueeze(0).expand(latent.shape[0], -1) + idx_mid = torch.arange(latents_history_mid.shape[2], device=latent.device, dtype=torch.int64).unsqueeze(0).expand(latent.shape[0], -1) + idx_long = torch.arange(latents_history_long.shape[2], device=latent.device, dtype=torch.int64).unsqueeze(0).expand(latent.shape[0], -1) + + values = { + "latents_history_short": latents_history_short, + "latents_history_mid": latents_history_mid, + "latents_history_long": latents_history_long, + "indices_latents_history_short": idx_short, + "indices_latents_history_mid": idx_mid, + "indices_latents_history_long": idx_long, + } + + positive = node_helpers.conditioning_set_values(positive, values) + negative = node_helpers.conditioning_set_values(negative, values) + return positive, negative + + +def _build_helios_indices(batch, history_sizes, keep_first_frame, hidden_frames, device): + sizes = list(history_sizes) + if len(sizes) != 3: + sizes = [16, 2, 1] + sizes = [max(0, int(v)) for v in sizes] + long_size, mid_size, short_base_size = sizes + + if keep_first_frame: + total = 1 + long_size + mid_size + short_base_size + hidden_frames + indices = torch.arange(total, device=device, dtype=torch.int64) + splits = [1, long_size, mid_size, short_base_size, hidden_frames] + indices_prefix, idx_long, idx_mid, idx_1x, idx_hidden = torch.split(indices, splits, dim=0) + idx_short = torch.cat([indices_prefix, idx_1x], dim=0) + else: + total = long_size + mid_size + short_base_size + hidden_frames + indices = torch.arange(total, device=device, dtype=torch.int64) + splits = [long_size, mid_size, short_base_size, hidden_frames] + idx_long, idx_mid, idx_short, idx_hidden = torch.split(indices, splits, dim=0) + + idx_hidden = idx_hidden.unsqueeze(0).expand(batch, -1) + idx_short = idx_short.unsqueeze(0).expand(batch, -1) + idx_mid = idx_mid.unsqueeze(0).expand(batch, -1) + idx_long = idx_long.unsqueeze(0).expand(batch, -1) + return idx_hidden, idx_short, idx_mid, idx_long + + +class HeliosImageToVideo(io.ComfyNode): + @classmethod + def define_schema(cls): + return io.Schema( + node_id="HeliosImageToVideo", + category="conditioning/video_models", + inputs=[ + io.Conditioning.Input("positive"), + io.Conditioning.Input("negative"), + io.Vae.Input("vae"), + io.Int.Input("width", default=640, min=16, max=nodes.MAX_RESOLUTION, step=16), + io.Int.Input("height", default=384, min=16, max=nodes.MAX_RESOLUTION, step=16), + io.Int.Input("length", default=132, min=1, max=nodes.MAX_RESOLUTION, step=4), + io.Int.Input("batch_size", default=1, min=1, max=4096), + io.Image.Input("start_image", optional=True), + io.String.Input("history_sizes", default="16,2,1", advanced=True), + io.Boolean.Input("keep_first_frame", default=True, advanced=True), + io.Int.Input("num_latent_frames_per_chunk", default=9, min=1, max=256, advanced=True), + io.Boolean.Input("add_noise_to_image_latents", default=True, advanced=True), + io.Float.Input("image_noise_sigma_min", default=0.111, min=0.0, max=1.0, step=0.0001, round=False, advanced=True), + io.Float.Input("image_noise_sigma_max", default=0.135, min=0.0, max=1.0, step=0.0001, round=False, advanced=True), + io.Int.Input("noise_seed", default=0, min=0, max=0xFFFFFFFFFFFFFFFF, advanced=True), + ], + outputs=[ + io.Conditioning.Output(display_name="positive"), + io.Conditioning.Output(display_name="negative"), + io.Latent.Output(display_name="latent"), + ], + ) + + @classmethod + def execute( + cls, + positive, + negative, + vae, + width, + height, + length, + batch_size, + start_image=None, + history_sizes="16,2,1", + keep_first_frame=True, + num_latent_frames_per_chunk=9, + add_noise_to_image_latents=True, + image_noise_sigma_min=0.111, + image_noise_sigma_max=0.135, + noise_seed=0, + ) -> io.NodeOutput: + video_noise_sigma_min = 0.111 + video_noise_sigma_max = 0.135 + spacial_scale = vae.spacial_compression_encode() + latent_channels = vae.latent_channels + latent_t = ((length - 1) // 4) + 1 + latent = torch.zeros([batch_size, latent_channels, latent_t, height // spacial_scale, width // spacial_scale], device=comfy.model_management.intermediate_device()) + + sizes = _parse_int_list(history_sizes, [16, 2, 1]) + if len(sizes) != 3: + sizes = [16, 2, 1] + sizes = sorted([max(0, int(v)) for v in sizes], reverse=True) + hist_len = max(1, sum(sizes)) + history_latent = torch.zeros([batch_size, latent_channels, hist_len, latent.shape[-2], latent.shape[-1]], device=latent.device, dtype=latent.dtype) + history_valid_mask = torch.zeros((batch_size, hist_len), device=latent.device, dtype=torch.bool) + image_latent_prefix = None + i2v_noise_gen = None + noise_gen_state = None + + if start_image is not None: + image = comfy.utils.common_upscale(start_image[:1].movedim(-1, 1), width, height, "bilinear", "center").movedim(1, -1) + img_latent = vae.encode(image[:, :, :, :3]).to(device=latent.device, dtype=torch.float32) + img_latent = comfy.utils.repeat_to_batch_size(img_latent, batch_size) + image_latent_prefix = img_latent[:, :, :1] + + if add_noise_to_image_latents: + i2v_noise_gen = torch.Generator(device=img_latent.device) + i2v_noise_gen.manual_seed(int(noise_seed)) + sigma = ( + torch.rand((1,), device=img_latent.device, generator=i2v_noise_gen, dtype=img_latent.dtype).view(1, 1, 1, 1, 1) + * (float(image_noise_sigma_max) - float(image_noise_sigma_min)) + + float(image_noise_sigma_min) + ) + image_latent_prefix = _apply_helios_latent_space_noise(image_latent_prefix, sigma, generator=i2v_noise_gen) + + min_frames = max(1, (int(num_latent_frames_per_chunk) - 1) * 4 + 1) + fake_video = image.repeat(min_frames, 1, 1, 1) + fake_latents_full = vae.encode(fake_video).to(device=latent.device, dtype=torch.float32) + fake_latent = comfy.utils.repeat_to_batch_size(fake_latents_full[:, :, -1:], batch_size) + # when adding noise to image latents, fake_image_latents used for history are also noised. + if add_noise_to_image_latents: + if i2v_noise_gen is None: + i2v_noise_gen = torch.Generator(device=fake_latent.device) + i2v_noise_gen.manual_seed(int(noise_seed)) + # Keep backward compatibility with existing I2V node inputs: + # this node exposes only image sigma controls; fake history latents + # follow the video-noise defaults. + fake_sigma = ( + torch.rand((1,), device=fake_latent.device, generator=i2v_noise_gen, dtype=fake_latent.dtype).view(1, 1, 1, 1, 1) + * (float(video_noise_sigma_max) - float(video_noise_sigma_min)) + + float(video_noise_sigma_min) + ) + fake_latent = _apply_helios_latent_space_noise(fake_latent, fake_sigma, generator=i2v_noise_gen) + history_latent[:, :, -1:] = fake_latent + history_valid_mask[:, -1] = True + if i2v_noise_gen is not None: + noise_gen_state = i2v_noise_gen.get_state().clone() + + positive, negative = _set_helios_history_values(positive, negative, history_latent, sizes, keep_first_frame, prefix_latent=image_latent_prefix) + return io.NodeOutput( + positive, + negative, + { + "samples": latent, + "helios_history_latent": history_latent, + "helios_image_latent_prefix": image_latent_prefix, + "helios_history_valid_mask": history_valid_mask, + "helios_num_frames": int(length), + "helios_noise_gen_state": noise_gen_state, + }, + ) + + +class HeliosTextToVideo(io.ComfyNode): + @classmethod + def define_schema(cls): + return io.Schema( + node_id="HeliosTextToVideo", + category="conditioning/video_models", + inputs=[ + io.Conditioning.Input("positive"), + io.Conditioning.Input("negative"), + io.Vae.Input("vae"), + io.Int.Input("width", default=640, min=16, max=nodes.MAX_RESOLUTION, step=16), + io.Int.Input("height", default=384, min=16, max=nodes.MAX_RESOLUTION, step=16), + io.Int.Input("length", default=132, min=1, max=nodes.MAX_RESOLUTION, step=4), + io.Int.Input("batch_size", default=1, min=1, max=4096), + io.String.Input("history_sizes", default="16,2,1", advanced=True), + ], + outputs=[ + io.Conditioning.Output(display_name="positive"), + io.Conditioning.Output(display_name="negative"), + io.Latent.Output(display_name="latent"), + ], + ) + + @classmethod + def execute( + cls, + positive, + negative, + vae, + width, + height, + length, + batch_size, + history_sizes="16,2,1", + ) -> io.NodeOutput: + spacial_scale = vae.spacial_compression_encode() + latent_channels = vae.latent_channels + latent_t = ((length - 1) // 4) + 1 + + # Create zero latent as shape placeholder (noise will be generated in sampler) + latent = torch.zeros( + [batch_size, latent_channels, latent_t, height // spacial_scale, width // spacial_scale], + device=comfy.model_management.intermediate_device(), + ) + + sizes = _parse_int_list(history_sizes, [16, 2, 1]) + if len(sizes) != 3: + sizes = [16, 2, 1] + sizes = sorted([max(0, int(v)) for v in sizes], reverse=True) + hist_len = max(1, sum(sizes)) + # History latent starts as zeros (no history yet) + history_latent = torch.zeros( + [batch_size, latent_channels, hist_len, latent.shape[-2], latent.shape[-1]], + device=latent.device, + dtype=latent.dtype, + ) + history_valid_mask = torch.zeros((batch_size, hist_len), device=latent.device, dtype=torch.bool) + + positive, negative = _set_helios_history_values( + positive, + negative, + history_latent, + sizes, + False, + prefix_latent=None, + ) + return io.NodeOutput( + positive, + negative, + { + "samples": latent, + "helios_history_latent": history_latent, + "helios_image_latent_prefix": None, + "helios_history_valid_mask": history_valid_mask, + "helios_num_frames": int(length), + }, + ) + + +class HeliosVideoToVideo(io.ComfyNode): + @classmethod + def define_schema(cls): + return io.Schema( + node_id="HeliosVideoToVideo", + category="conditioning/video_models", + inputs=[ + io.Conditioning.Input("positive"), + io.Conditioning.Input("negative"), + io.Vae.Input("vae"), + io.Int.Input("width", default=640, min=16, max=nodes.MAX_RESOLUTION, step=16), + io.Int.Input("height", default=384, min=16, max=nodes.MAX_RESOLUTION, step=16), + io.Int.Input("length", default=132, min=1, max=nodes.MAX_RESOLUTION, step=4), + io.Int.Input("batch_size", default=1, min=1, max=4096), + io.Image.Input("video", optional=True), + io.String.Input("history_sizes", default="16,2,1", advanced=True), + io.Boolean.Input("keep_first_frame", default=True, advanced=True), + io.Int.Input("num_latent_frames_per_chunk", default=9, min=1, max=256, advanced=True), + io.Boolean.Input("add_noise_to_video_latents", default=True, advanced=True), + io.Float.Input("video_noise_sigma_min", default=0.111, min=0.0, max=1.0, step=0.0001, round=False, advanced=True), + io.Float.Input("video_noise_sigma_max", default=0.135, min=0.0, max=1.0, step=0.0001, round=False, advanced=True), + io.Int.Input("noise_seed", default=0, min=0, max=0xFFFFFFFFFFFFFFFF, advanced=True), + ], + outputs=[ + io.Conditioning.Output(display_name="positive"), + io.Conditioning.Output(display_name="negative"), + io.Latent.Output(display_name="latent"), + ], + ) + + @classmethod + def execute( + cls, + positive, + negative, + vae, + width, + height, + length, + batch_size, + video=None, + history_sizes="16,2,1", + keep_first_frame=True, + num_latent_frames_per_chunk=9, + add_noise_to_video_latents=True, + video_noise_sigma_min=0.111, + video_noise_sigma_max=0.135, + noise_seed=0, + ) -> io.NodeOutput: + spacial_scale = vae.spacial_compression_encode() + latent_channels = vae.latent_channels + latent_t = ((length - 1) // 4) + 1 + latent = torch.zeros([batch_size, latent_channels, latent_t, height // spacial_scale, width // spacial_scale], device=comfy.model_management.intermediate_device()) + + sizes = _parse_int_list(history_sizes, [16, 2, 1]) + if len(sizes) != 3: + sizes = [16, 2, 1] + sizes = sorted([max(0, int(v)) for v in sizes], reverse=True) + hist_len = max(1, sum(sizes)) + history_latent = torch.zeros([batch_size, latent_channels, hist_len, latent.shape[-2], latent.shape[-1]], device=latent.device, dtype=latent.dtype) + history_valid_mask = torch.zeros((batch_size, hist_len), device=latent.device, dtype=torch.bool) + image_latent_prefix = None + noise_gen_state = None + history_latent_output = history_latent + + if video is not None: + video = comfy.utils.common_upscale(video[:length].movedim(-1, 1), width, height, "bilinear", "center").movedim(1, -1) + num_frames = int(video.shape[0]) + min_frames = max(1, (int(num_latent_frames_per_chunk) - 1) * 4 + 1) + num_chunks = num_frames // min_frames + if num_chunks == 0: + raise ValueError( + f"Video must have at least {min_frames} frames (got {num_frames} frames). " + f"Required: (num_latent_frames_per_chunk - 1) * 4 + 1 = ({int(num_latent_frames_per_chunk)} - 1) * 4 + 1 = {min_frames}" + ) + + first_frame = video[:1] + first_frame_latent = vae.encode(first_frame[:, :, :, :3]).to(device=latent.device, dtype=torch.float32) + + total_valid_frames = num_chunks * min_frames + start_frame = num_frames - total_valid_frames + latents_chunks = [] + for i in range(num_chunks): + chunk_start = start_frame + i * min_frames + chunk_end = chunk_start + min_frames + video_chunk = video[chunk_start:chunk_end] + chunk_latents = vae.encode(video_chunk[:, :, :, :3]).to(device=latent.device, dtype=torch.float32) + latents_chunks.append(chunk_latents) + vid_latent = torch.cat(latents_chunks, dim=2) + vid_latent_clean = vid_latent.clone() + + if add_noise_to_video_latents: + g = torch.Generator(device=vid_latent.device) + g.manual_seed(int(noise_seed)) + + image_sigma = ( + torch.rand((1,), device=first_frame_latent.device, generator=g, dtype=first_frame_latent.dtype).view(1, 1, 1, 1, 1) + * (float(video_noise_sigma_max) - float(video_noise_sigma_min)) + + float(video_noise_sigma_min) + ) + first_frame_latent = _apply_helios_latent_space_noise(first_frame_latent, image_sigma, generator=g) + + noisy_chunks = [] + num_latent_chunks = max(1, vid_latent.shape[2] // int(num_latent_frames_per_chunk)) + for i in range(num_latent_chunks): + chunk_start = i * int(num_latent_frames_per_chunk) + chunk_end = chunk_start + int(num_latent_frames_per_chunk) + latent_chunk = vid_latent[:, :, chunk_start:chunk_end, :, :] + if latent_chunk.shape[2] == 0: + continue + chunk_frames = latent_chunk.shape[2] + frame_sigmas = ( + torch.rand((chunk_frames,), device=vid_latent.device, generator=g, dtype=vid_latent.dtype) + * (float(video_noise_sigma_max) - float(video_noise_sigma_min)) + + float(video_noise_sigma_min) + ).view(1, 1, chunk_frames, 1, 1) + noisy_chunk = _apply_helios_latent_space_noise(latent_chunk, frame_sigmas, generator=g) + noisy_chunks.append(noisy_chunk) + if len(noisy_chunks) > 0: + vid_latent = torch.cat(noisy_chunks, dim=2) + noise_gen_state = g.get_state().clone() + vid_latent = comfy.utils.repeat_to_batch_size(vid_latent, batch_size) + image_latent_prefix = comfy.utils.repeat_to_batch_size(first_frame_latent, batch_size) + video_frames = vid_latent.shape[2] + if video_frames < hist_len: + keep_frames = hist_len - video_frames + history_latent = torch.cat([history_latent[:, :, :keep_frames], vid_latent], dim=2) + history_latent_output = torch.cat([history_latent_output[:, :, :keep_frames], comfy.utils.repeat_to_batch_size(vid_latent_clean, batch_size)], dim=2) + history_valid_mask[:, keep_frames:] = True + else: + history_latent = vid_latent + history_latent_output = comfy.utils.repeat_to_batch_size(vid_latent_clean, batch_size) + history_valid_mask = torch.ones((batch_size, video_frames), device=latent.device, dtype=torch.bool) + + positive, negative = _set_helios_history_values(positive, negative, history_latent, sizes, keep_first_frame, prefix_latent=image_latent_prefix) + return io.NodeOutput( + positive, + negative, + { + "samples": latent, + "helios_history_latent": history_latent, + "helios_history_latent_output": history_latent_output, + "helios_image_latent_prefix": image_latent_prefix, + "helios_history_valid_mask": history_valid_mask, + "helios_num_frames": int(length), + "helios_noise_gen_state": noise_gen_state, + }, + ) + + +class HeliosPyramidSampler(io.ComfyNode): + @classmethod + def define_schema(cls): + return io.Schema( + node_id="HeliosPyramidSampler", + category="sampling/video_models", + inputs=[ + io.Model.Input("model"), + io.Int.Input("noise_seed", default=0, min=0, max=0xFFFFFFFFFFFFFFFF, control_after_generate=True), + io.Float.Input("cfg", default=5.0, min=0.0, max=100.0, step=0.1, round=0.01), + io.Conditioning.Input("positive"), + io.Conditioning.Input("negative"), + io.Latent.Input("latent_image"), + io.String.Input("pyramid_steps", default="10,10,10"), + io.String.Input("stage_range", default="0,0.333333,0.666667,1"), + io.Boolean.Input("distilled", default=False), + io.Boolean.Input("amplify_first_stage", default=False), + io.Float.Input("gamma", default=1.0 / 3.0, min=0.0001, max=10.0, step=0.0001, round=False), + io.String.Input("history_sizes", default="16,2,1", advanced=True), + io.Boolean.Input("keep_first_frame", default=True, advanced=True), + io.Int.Input("num_latent_frames_per_chunk", default=9, min=1, max=256, advanced=True), + io.Boolean.Input("cfg_zero_star", default=True, advanced=True), + io.Boolean.Input("use_zero_init", default=True, advanced=True), + io.Int.Input("zero_steps", default=1, min=0, max=10000, advanced=True), + ], + outputs=[ + io.Latent.Output(display_name="output"), + ], + ) + + @classmethod + def execute( + cls, + model, + noise_seed, + cfg, + positive, + negative, + latent_image, + pyramid_steps, + stage_range, + distilled, + amplify_first_stage, + gamma, + history_sizes, + keep_first_frame, + num_latent_frames_per_chunk, + cfg_zero_star, + use_zero_init, + zero_steps, + ) -> io.NodeOutput: + # Keep these scheduler knobs internal (not exposed in node UI). + shift = 1.0 + num_train_timesteps = 1000 + # Keep dynamic shifting always on for Helios parity; not exposed in node UI. + use_dynamic_shifting = True + time_shift_type = "exponential" + base_image_seq_len = 256 + max_image_seq_len = 4096 + base_shift = 0.5 + max_shift = 1.15 + + latent = latent_image.copy() + latent_samples = comfy.sample.fix_empty_latent_channels(model, latent["samples"], latent.get("downscale_ratio_spacial", None)) + + stage_steps = _parse_int_list(pyramid_steps, [10, 10, 10]) + stage_steps = [max(1, int(s)) for s in stage_steps] + stage_count = len(stage_steps) + history_sizes_list = sorted([max(0, int(v)) for v in _parse_int_list(history_sizes, [16, 2, 1])], reverse=True) + if not keep_first_frame and len(history_sizes_list) > 0: + history_sizes_list[-1] += 1 + + stage_range_values = _parse_float_list(stage_range, [0.0, 1.0 / 3.0, 2.0 / 3.0, 1.0]) + if len(stage_range_values) != stage_count + 1: + stage_range_values = [float(i) / float(stage_count) for i in range(stage_count + 1)] + + stage_tables = _helios_stage_tables( + stage_count=stage_count, + stage_range=stage_range_values, + gamma=float(gamma), + num_train_timesteps=int(num_train_timesteps), + shift=float(shift), + ) + + b, c, t, h, w = latent_samples.shape + chunk_t = max(1, int(num_latent_frames_per_chunk)) + num_frames = int(latent.get("helios_num_frames", max(1, (int(t) - 1) * 4 + 1))) + window_num_frames = (chunk_t - 1) * 4 + 1 + chunk_count = max(1, (num_frames + window_num_frames - 1) // window_num_frames) + euler_sampler = comfy.samplers.KSAMPLER(_helios_euler_sample) + target_device = comfy.model_management.get_torch_device() + noise_gen = torch.Generator(device=target_device) + noise_gen.manual_seed(int(noise_seed)) + noise_gen_state = latent.get("helios_noise_gen_state", None) + if noise_gen_state is not None: + try: + noise_gen.set_state(noise_gen_state) + except Exception: + pass + + image_latent_prefix = latent.get("helios_image_latent_prefix", None) + history_valid_mask = latent.get("helios_history_valid_mask", None) + if history_valid_mask is None: + raise ValueError("Helios sampler requires `helios_history_valid_mask` in latent input.") + history_from_latent_applied = False + if image_latent_prefix is not None: + image_latent_prefix = model.model.process_latent_in(image_latent_prefix) + if "helios_history_latent" in latent: + history_in = _process_latent_in_preserve_zero_frames(model, latent["helios_history_latent"], valid_mask=history_valid_mask) + positive, negative = _set_helios_history_values( + positive, + negative, + history_in, + history_sizes_list, + keep_first_frame, + prefix_latent=image_latent_prefix, + ) + history_from_latent_applied = True + + latents_history_short = _extract_condition_value(positive, "latents_history_short") + latents_history_mid = _extract_condition_value(positive, "latents_history_mid") + latents_history_long = _extract_condition_value(positive, "latents_history_long") + if (not history_from_latent_applied) and latents_history_short is not None and latents_history_mid is not None and latents_history_long is not None: + raise ValueError("Helios requires `helios_history_latent` + `helios_history_valid_mask`; direct history conditioning is not supported.") + if latents_history_short is None and "helios_history_latent" in latent: + history_in = _process_latent_in_preserve_zero_frames(model, latent["helios_history_latent"], valid_mask=history_valid_mask) + positive, negative = _set_helios_history_values( + positive, + negative, + history_in, + history_sizes_list, + keep_first_frame, + prefix_latent=image_latent_prefix, + ) + latents_history_short = _extract_condition_value(positive, "latents_history_short") + latents_history_mid = _extract_condition_value(positive, "latents_history_mid") + latents_history_long = _extract_condition_value(positive, "latents_history_long") + + x0_output = {} + generated_chunks = [] + if latents_history_short is not None and latents_history_mid is not None and latents_history_long is not None: + short_base_size = history_sizes_list[-1] if len(history_sizes_list) > 0 else latents_history_short.shape[2] + if keep_first_frame and latents_history_short.shape[2] > short_base_size: + short_for_history = latents_history_short[:, :, -short_base_size:] + else: + short_for_history = latents_history_short + rolling_history = torch.cat([latents_history_long, latents_history_mid, short_for_history], dim=2) + elif "helios_history_latent" in latent: + rolling_history = latent["helios_history_latent"] + rolling_history = _process_latent_in_preserve_zero_frames(model, rolling_history, valid_mask=history_valid_mask) + else: + hist_len = max(1, sum(history_sizes_list)) + rolling_history = torch.zeros((b, c, hist_len, h, w), device=latent_samples.device, dtype=latent_samples.dtype) + + # Keep history/prefix on the same device/dtype as denoising latents. + rolling_history = rolling_history.to(device=target_device, dtype=torch.float32) + if image_latent_prefix is not None: + image_latent_prefix = image_latent_prefix.to(device=target_device, dtype=torch.float32) + + # Always return only newly generated chunks; input history is used only for conditioning. + + for chunk_idx in range(chunk_count): + # Prepare initial latent for this chunk + noise_shape = ( + latent_samples.shape[0], + latent_samples.shape[1], + chunk_t, + latent_samples.shape[3], + latent_samples.shape[4], + ) + stage_latent = torch.randn(noise_shape, device=target_device, dtype=torch.float32, generator=noise_gen) + + # Downsample to stage 0 resolution + for _ in range(max(0, int(stage_count) - 1)): + stage_latent = _downsample_latent_5d_bilinear_x2(stage_latent) + + # Keep stage latents on model device for scheduler/noise path consistency. + stage_latent = stage_latent.to(target_device) + + chunk_prefix = image_latent_prefix + if keep_first_frame and image_latent_prefix is None and chunk_idx == 0: + chunk_prefix = torch.zeros( + ( + rolling_history.shape[0], + rolling_history.shape[1], + 1, + rolling_history.shape[3], + rolling_history.shape[4], + ), + device=rolling_history.device, + dtype=rolling_history.dtype, + ) + + positive_chunk, negative_chunk = _set_helios_history_values( + positive, + negative, + rolling_history, + history_sizes_list, + keep_first_frame, + prefix_latent=chunk_prefix, + ) + latents_history_short = _extract_condition_value(positive_chunk, "latents_history_short") + latents_history_mid = _extract_condition_value(positive_chunk, "latents_history_mid") + latents_history_long = _extract_condition_value(positive_chunk, "latents_history_long") + for stage_idx in range(stage_count): + stage_latent = stage_latent.to(comfy.model_management.get_torch_device()) + sigmas = _helios_stage_sigmas( + stage_idx=stage_idx, + stage_steps=stage_steps[stage_idx], + stage_tables=stage_tables, + is_distilled=distilled, + is_amplify_first_stage=amplify_first_stage and chunk_idx == 0, + ).to(device=stage_latent.device, dtype=torch.float32) + timesteps = _helios_stage_timesteps( + stage_idx=stage_idx, + stage_steps=stage_steps[stage_idx], + stage_tables=stage_tables, + is_distilled=distilled, + is_amplify_first_stage=amplify_first_stage and chunk_idx == 0, + ).to(device=stage_latent.device, dtype=torch.float32) + if use_dynamic_shifting: + patch_size = (1, 2, 2) + image_seq_len = (stage_latent.shape[-1] * stage_latent.shape[-2] * stage_latent.shape[-3]) // (patch_size[0] * patch_size[1] * patch_size[2]) + mu = _calculate_shift( + image_seq_len=image_seq_len, + base_seq_len=base_image_seq_len, + max_seq_len=max_image_seq_len, + base_shift=base_shift, + max_shift=max_shift, + ) + sigmas = _time_shift(sigmas, mu=mu, sigma=1.0, mode=time_shift_type).to(torch.float32) + tmin = torch.min(timesteps) + tmax = torch.max(timesteps) + timesteps = tmin + sigmas[:-1] * (tmax - tmin) + else: + pass + + # Stage timesteps are computed before upsampling/renoise for stage > 0. + if stage_idx > 0: + stage_latent = _upsample_latent_5d(stage_latent, scale=2) + + ori_sigma = 1.0 - float(stage_tables["ori_start_sigmas"][stage_idx]) + alpha = 1.0 / (math.sqrt(1.0 + (1.0 / gamma)) * (1.0 - ori_sigma) + ori_sigma) + beta = alpha * (1.0 - ori_sigma) / math.sqrt(gamma) + + noise = _sample_block_noise_like(stage_latent, gamma, patch_size=(1, 2, 2), generator=noise_gen).to(stage_latent) + stage_latent = alpha * stage_latent + beta * noise + + indices_hidden_states, idx_short, idx_mid, idx_long = _build_helios_indices( + batch=stage_latent.shape[0], + history_sizes=history_sizes_list, + keep_first_frame=keep_first_frame, + hidden_frames=stage_latent.shape[2], + device=stage_latent.device, + ) + positive_stage = node_helpers.conditioning_set_values(positive_chunk, {"indices_hidden_states": indices_hidden_states}) + negative_stage = node_helpers.conditioning_set_values(negative_chunk, {"indices_hidden_states": indices_hidden_states}) + + if latents_history_short is not None: + values = {"latents_history_short": latents_history_short, "indices_latents_history_short": idx_short} + positive_stage = node_helpers.conditioning_set_values(positive_stage, values) + negative_stage = node_helpers.conditioning_set_values(negative_stage, values) + + if latents_history_mid is not None: + values = {"latents_history_mid": latents_history_mid, "indices_latents_history_mid": idx_mid} + positive_stage = node_helpers.conditioning_set_values(positive_stage, values) + negative_stage = node_helpers.conditioning_set_values(negative_stage, values) + + if latents_history_long is not None: + values = {"latents_history_long": latents_history_long, "indices_latents_history_long": idx_long} + positive_stage = node_helpers.conditioning_set_values(positive_stage, values) + negative_stage = node_helpers.conditioning_set_values(negative_stage, values) + + stage_time_values = { + "helios_stage_sigmas": sigmas, + "helios_stage_timesteps": timesteps, + } + positive_stage = node_helpers.conditioning_set_values(positive_stage, stage_time_values) + negative_stage = node_helpers.conditioning_set_values(negative_stage, stage_time_values) + + cfg_use = 1.0 if distilled else cfg + + sigma0 = max(float(sigmas[0].item()), 1e-6) + noise = stage_latent / sigma0 + latent_start = torch.zeros_like(stage_latent) + + stage_start_for_dmd = stage_latent.clone() + + if distilled: + sampler = comfy.samplers.KSAMPLER( + _helios_dmd_sample, + extra_options={ + "dmd_noisy_tensor": stage_start_for_dmd, + "dmd_sigmas": sigmas, + "dmd_timesteps": timesteps, + "all_timesteps": timesteps, + }, + ) + else: + sampler = euler_sampler + + callback = latent_preview.prepare_callback(model, sigmas.shape[-1] - 1, x0_output) + stage_model = model + if cfg_zero_star and not distilled: + stage_model = model.clone() + stage_model.model_options = comfy.model_patcher.set_model_options_pre_cfg_function( + stage_model.model_options, + _build_cfg_zero_star_pre_cfg(stage_idx=stage_idx, zero_steps=zero_steps, use_zero_init=use_zero_init), + disable_cfg1_optimization=True, + ) + stage_latent = comfy.sample.sample_custom( + stage_model, + noise, + cfg_use, + sampler, + sigmas, + positive_stage, + negative_stage, + latent_start, + noise_mask=None, + callback=callback, + disable_pbar=not comfy.utils.PROGRESS_BAR_ENABLED, + seed=noise_seed + chunk_idx * 100 + stage_idx, + ) + # sample_custom returns latent_format.process_out(samples); convert back to model-space + # so subsequent pyramid stages and history conditioning stay in the same latent space. + stage_latent = model.model.process_latent_in(stage_latent) + + if stage_latent.shape[-2] != h or stage_latent.shape[-1] != w: + b2, c2, t2, h2, w2 = stage_latent.shape + x = stage_latent.permute(0, 2, 1, 3, 4).reshape(b2 * t2, c2, h2, w2) + x = comfy.utils.common_upscale(x, w, h, "nearest-exact", "disabled") + stage_latent = x.reshape(b2, t2, c2, h, w).permute(0, 2, 1, 3, 4) + stage_latent = stage_latent[:, :, :, :h, :w] + + generated_chunks.append(stage_latent) + if keep_first_frame and (chunk_idx == 0 and image_latent_prefix is None): + image_latent_prefix = stage_latent[:, :, :1] + rolling_history = torch.cat([rolling_history, stage_latent.to(rolling_history.device, rolling_history.dtype)], dim=2) + keep_hist = max(1, sum(history_sizes_list)) + rolling_history = rolling_history[:, :, -keep_hist:] + + if len(generated_chunks) > 0: + stage_latent = torch.cat(generated_chunks, dim=2) + else: + stage_latent = torch.zeros((b, c, 0, h, w), device=target_device, dtype=torch.float32) + + out = latent.copy() + out.pop("downscale_ratio_spacial", None) + out["samples"] = model.model.process_latent_out(stage_latent) + out["helios_chunk_decode"] = True + out["helios_chunk_latent_frames"] = int(chunk_t) + out["helios_chunk_count"] = int(len(generated_chunks)) + out["helios_window_num_frames"] = int(window_num_frames) + out["helios_num_frames"] = int(num_frames) + + return io.NodeOutput(out) + + +class HeliosVAEDecode(io.ComfyNode): + @classmethod + def define_schema(cls): + return io.Schema( + node_id="HeliosVAEDecode", + category="latent", + inputs=[ + io.Latent.Input("samples"), + io.Vae.Input("vae"), + ], + outputs=[io.Image.Output(display_name="image")], + ) + + @classmethod + def execute(cls, samples, vae) -> io.NodeOutput: + latent = samples["samples"] + if latent.is_nested: + latent = latent.unbind()[0] + + helios_chunk_decode = bool(samples.get("helios_chunk_decode", False)) + helios_chunk_latent_frames = int(samples.get("helios_chunk_latent_frames", 0) or 0) + + if ( + helios_chunk_decode + and latent.ndim == 5 + and helios_chunk_latent_frames > 0 + and latent.shape[2] > 0 + ): + decoded_chunks = [] + body = latent + for start in range(0, body.shape[2], helios_chunk_latent_frames): + chunk = body[:, :, start:start + helios_chunk_latent_frames] + if chunk.shape[2] == 0: + continue + decoded_chunks.append(vae.decode(chunk)) + + if len(decoded_chunks) > 0: + images = torch.cat(decoded_chunks, dim=1) + else: + images = vae.decode(latent) + else: + images = vae.decode(latent) + + if len(images.shape) == 5: + images = images.reshape(-1, images.shape[-3], images.shape[-2], images.shape[-1]) + return io.NodeOutput(images) + + +class HeliosExtension(ComfyExtension): + @override + async def get_node_list(self) -> list[type[io.ComfyNode]]: + return [ + HeliosTextToVideo, + HeliosImageToVideo, + HeliosVideoToVideo, + HeliosPyramidSampler, + HeliosVAEDecode, + ] + + +async def comfy_entrypoint() -> HeliosExtension: + return HeliosExtension() diff --git a/nodes.py b/nodes.py index 1e19a8223..c73bac2c1 100644 --- a/nodes.py +++ b/nodes.py @@ -976,7 +976,7 @@ class CLIPLoader: @classmethod def INPUT_TYPES(s): return {"required": { "clip_name": (folder_paths.get_filename_list("text_encoders"), ), - "type": (["stable_diffusion", "stable_cascade", "sd3", "stable_audio", "mochi", "ltxv", "pixart", "cosmos", "lumina2", "wan", "hidream", "chroma", "ace", "omnigen2", "qwen_image", "hunyuan_image", "flux2", "ovis", "longcat_image"], ), + "type": (["stable_diffusion", "stable_cascade", "sd3", "stable_audio", "mochi", "ltxv", "pixart", "cosmos", "lumina2", "wan", "helios", "hidream", "chroma", "ace", "omnigen2", "qwen_image", "hunyuan_image", "flux2", "ovis", "longcat_image"], ), }, "optional": { "device": (["default", "cpu"], {"advanced": True}), @@ -986,7 +986,7 @@ class CLIPLoader: CATEGORY = "advanced/loaders" - DESCRIPTION = "[Recipes]\n\nstable_diffusion: clip-l\nstable_cascade: clip-g\nsd3: t5 xxl/ clip-g / clip-l\nstable_audio: t5 base\nmochi: t5 xxl\ncosmos: old t5 xxl\nlumina2: gemma 2 2B\nwan: umt5 xxl\n hidream: llama-3.1 (Recommend) or t5\nomnigen2: qwen vl 2.5 3B" + DESCRIPTION = "[Recipes]\n\nstable_diffusion: clip-l\nstable_cascade: clip-g\nsd3: t5 xxl/ clip-g / clip-l\nstable_audio: t5 base\nmochi: t5 xxl\ncosmos: old t5 xxl\nlumina2: gemma 2 2B\nwan: umt5 xxl\nhelios: umt5 xxl\n hidream: llama-3.1 (Recommend) or t5\nomnigen2: qwen vl 2.5 3B" def load_clip(self, clip_name, type="stable_diffusion", device="default"): clip_type = getattr(comfy.sd.CLIPType, type.upper(), comfy.sd.CLIPType.STABLE_DIFFUSION) @@ -2414,6 +2414,7 @@ async def init_builtin_extra_nodes(): "nodes_cosmos.py", "nodes_video.py", "nodes_lumina2.py", + "nodes_helios.py", "nodes_wan.py", "nodes_lotus.py", "nodes_hunyuan3d.py",