2024-06-26 15:38:02 +00:00
|
|
|
from contextlib import nullcontext
|
2024-06-05 17:47:38 +00:00
|
|
|
from functools import singledispatchmethod
|
|
|
|
|
|
|
|
import einops
|
|
|
|
import torch
|
|
|
|
from diffusers.models.attention_processor import (
|
|
|
|
AttnProcessor2_0,
|
|
|
|
LoRAAttnProcessor2_0,
|
|
|
|
LoRAXFormersAttnProcessor,
|
|
|
|
XFormersAttnProcessor,
|
|
|
|
)
|
|
|
|
from diffusers.models.autoencoders.autoencoder_kl import AutoencoderKL
|
|
|
|
from diffusers.models.autoencoders.autoencoder_tiny import AutoencoderTiny
|
|
|
|
|
|
|
|
from invokeai.app.invocations.baseinvocation import BaseInvocation, invocation
|
2024-06-28 15:30:09 +00:00
|
|
|
from invokeai.app.invocations.constants import DEFAULT_PRECISION, LATENT_SCALE_FACTOR
|
2024-06-05 17:47:38 +00:00
|
|
|
from invokeai.app.invocations.fields import (
|
|
|
|
FieldDescriptions,
|
|
|
|
ImageField,
|
|
|
|
Input,
|
|
|
|
InputField,
|
|
|
|
)
|
|
|
|
from invokeai.app.invocations.model import VAEField
|
|
|
|
from invokeai.app.invocations.primitives import LatentsOutput
|
|
|
|
from invokeai.app.services.shared.invocation_context import InvocationContext
|
|
|
|
from invokeai.backend.model_manager import LoadedModel
|
|
|
|
from invokeai.backend.stable_diffusion.diffusers_pipeline import image_resized_to_grid_as_tensor
|
2024-06-26 15:38:02 +00:00
|
|
|
from invokeai.backend.stable_diffusion.vae_tiling import patch_vae_tiling_params
|
2024-06-05 17:47:38 +00:00
|
|
|
|
|
|
|
|
|
|
|
@invocation(
|
|
|
|
"i2l",
|
|
|
|
title="Image to Latents",
|
|
|
|
tags=["latents", "image", "vae", "i2l"],
|
|
|
|
category="latents",
|
2024-06-28 15:30:09 +00:00
|
|
|
version="1.1.0",
|
2024-06-05 17:47:38 +00:00
|
|
|
)
|
|
|
|
class ImageToLatentsInvocation(BaseInvocation):
|
|
|
|
"""Encodes an image into latents."""
|
|
|
|
|
|
|
|
image: ImageField = InputField(
|
|
|
|
description="The image to encode",
|
|
|
|
)
|
|
|
|
vae: VAEField = InputField(
|
|
|
|
description=FieldDescriptions.vae,
|
|
|
|
input=Input.Connection,
|
|
|
|
)
|
|
|
|
tiled: bool = InputField(default=False, description=FieldDescriptions.tiled)
|
2024-06-28 15:30:09 +00:00
|
|
|
# NOTE: tile_size = 0 is a special value. We use this rather than `int | None`, because the workflow UI does not
|
|
|
|
# offer a way to directly set None values.
|
|
|
|
tile_size: int = InputField(default=0, multiple_of=8, description=FieldDescriptions.vae_tile_size)
|
2024-06-06 13:16:45 +00:00
|
|
|
fp32: bool = InputField(default=DEFAULT_PRECISION == torch.float32, description=FieldDescriptions.fp32)
|
2024-06-05 17:47:38 +00:00
|
|
|
|
|
|
|
@staticmethod
|
2024-06-28 15:30:09 +00:00
|
|
|
def vae_encode(
|
|
|
|
vae_info: LoadedModel, upcast: bool, tiled: bool, image_tensor: torch.Tensor, tile_size: int = 0
|
|
|
|
) -> torch.Tensor:
|
2024-06-05 17:47:38 +00:00
|
|
|
with vae_info as vae:
|
2024-06-26 15:38:02 +00:00
|
|
|
assert isinstance(vae, (AutoencoderKL, AutoencoderTiny))
|
2024-06-05 17:47:38 +00:00
|
|
|
orig_dtype = vae.dtype
|
|
|
|
if upcast:
|
|
|
|
vae.to(dtype=torch.float32)
|
|
|
|
|
|
|
|
use_torch_2_0_or_xformers = hasattr(vae.decoder, "mid_block") and isinstance(
|
|
|
|
vae.decoder.mid_block.attentions[0].processor,
|
|
|
|
(
|
|
|
|
AttnProcessor2_0,
|
|
|
|
XFormersAttnProcessor,
|
|
|
|
LoRAXFormersAttnProcessor,
|
|
|
|
LoRAAttnProcessor2_0,
|
|
|
|
),
|
|
|
|
)
|
|
|
|
# if xformers or torch_2_0 is used attention block does not need
|
|
|
|
# to be in float32 which can save lots of memory
|
|
|
|
if use_torch_2_0_or_xformers:
|
|
|
|
vae.post_quant_conv.to(orig_dtype)
|
|
|
|
vae.decoder.conv_in.to(orig_dtype)
|
|
|
|
vae.decoder.mid_block.to(orig_dtype)
|
|
|
|
# else:
|
|
|
|
# latents = latents.float()
|
|
|
|
|
|
|
|
else:
|
|
|
|
vae.to(dtype=torch.float16)
|
|
|
|
# latents = latents.half()
|
|
|
|
|
|
|
|
if tiled:
|
2024-06-28 15:30:09 +00:00
|
|
|
vae.enable_tiling()
|
|
|
|
else:
|
|
|
|
vae.disable_tiling()
|
|
|
|
|
|
|
|
tiling_context = nullcontext()
|
|
|
|
if tile_size > 0:
|
2024-06-26 15:38:02 +00:00
|
|
|
tiling_context = patch_vae_tiling_params(
|
|
|
|
vae,
|
2024-06-28 15:30:09 +00:00
|
|
|
tile_sample_min_size=tile_size,
|
|
|
|
tile_latent_min_size=tile_size // LATENT_SCALE_FACTOR,
|
2024-06-26 15:38:02 +00:00
|
|
|
tile_overlap_factor=0.25,
|
|
|
|
)
|
2024-06-05 17:47:38 +00:00
|
|
|
|
|
|
|
# non_noised_latents_from_image
|
|
|
|
image_tensor = image_tensor.to(device=vae.device, dtype=vae.dtype)
|
2024-06-26 15:38:02 +00:00
|
|
|
with torch.inference_mode(), tiling_context:
|
2024-06-05 17:47:38 +00:00
|
|
|
latents = ImageToLatentsInvocation._encode_to_tensor(vae, image_tensor)
|
|
|
|
|
|
|
|
latents = vae.config.scaling_factor * latents
|
|
|
|
latents = latents.to(dtype=orig_dtype)
|
|
|
|
|
|
|
|
return latents
|
|
|
|
|
|
|
|
@torch.no_grad()
|
|
|
|
def invoke(self, context: InvocationContext) -> LatentsOutput:
|
|
|
|
image = context.images.get_pil(self.image.image_name)
|
|
|
|
|
|
|
|
vae_info = context.models.load(self.vae.vae)
|
|
|
|
|
|
|
|
image_tensor = image_resized_to_grid_as_tensor(image.convert("RGB"))
|
|
|
|
if image_tensor.dim() == 3:
|
|
|
|
image_tensor = einops.rearrange(image_tensor, "c h w -> 1 c h w")
|
|
|
|
|
2024-06-28 15:30:09 +00:00
|
|
|
latents = self.vae_encode(
|
|
|
|
vae_info=vae_info, upcast=self.fp32, tiled=self.tiled, image_tensor=image_tensor, tile_size=self.tile_size
|
|
|
|
)
|
2024-06-05 17:47:38 +00:00
|
|
|
|
|
|
|
latents = latents.to("cpu")
|
|
|
|
name = context.tensors.save(tensor=latents)
|
|
|
|
return LatentsOutput.build(latents_name=name, latents=latents, seed=None)
|
|
|
|
|
|
|
|
@singledispatchmethod
|
|
|
|
@staticmethod
|
|
|
|
def _encode_to_tensor(vae: AutoencoderKL, image_tensor: torch.FloatTensor) -> torch.FloatTensor:
|
|
|
|
assert isinstance(vae, torch.nn.Module)
|
|
|
|
image_tensor_dist = vae.encode(image_tensor).latent_dist
|
|
|
|
latents: torch.Tensor = image_tensor_dist.sample().to(
|
|
|
|
dtype=vae.dtype
|
|
|
|
) # FIXME: uses torch.randn. make reproducible!
|
|
|
|
return latents
|
|
|
|
|
|
|
|
@_encode_to_tensor.register
|
|
|
|
@staticmethod
|
|
|
|
def _(vae: AutoencoderTiny, image_tensor: torch.FloatTensor) -> torch.FloatTensor:
|
|
|
|
assert isinstance(vae, torch.nn.Module)
|
|
|
|
latents: torch.FloatTensor = vae.encode(image_tensor).latents
|
|
|
|
return latents
|