1.stable_diffusion_controlnet_img2img.py or inpainting_img2img
主要区别在于latents的初始化方式
其中, inpainting: 输入变量image为
masked_image = image * (mask_image < 0.5)
def prepare_latents(self, image, timestep, batch_size, num_images_per_prompt, dtype, device, generator=None):
if not isinstance(image, (torch.Tensor, PIL.Image.Image, list)):
raise ValueError(
f"`image` has to be of type `torch.Tensor`, `PIL.Image.Image` or list but is {type(image)}"
)
image = image.to(device=device, dtype=dtype)
batch_size = batch_size * num_images_per_prompt
if isinstance(generator, list) and len(generator) != batch_size:
raise ValueError(
f"You have passed a list of generators of length {len(generator)}, but requested an effective batch"
f" size of {batch_size}. Make sure the batch size matches the length of the generators."
)
if isinstance(generator, list):
init_latents = [
self.vae.encode(image[i : i + 1]).latent_dist.sample(generator[i]) for i in range(batch_size)
]
init_latents = torch.cat(init_latents, dim=0)
else:
init_latents = self.vae.encode(image).latent_dist.sample(generator)
init_latents = self.vae.config.scaling_factor * init_latents
if batch_size > init_latents.shape[0] and batch_size % init_latents.shape[0] == 0:
raise ValueError(
f"Cannot duplicate `image` of batch size {init_latents.shape[0]} to {batch_size} text prompts."
)
else:
init_latents = torch.cat([init_latents], dim=0)
shape = init_latents.shape
noise = randn_tensor(shape, generator=generator, device=device, dtype=dtype)
# get latents
init_latents = self.scheduler.add_noise(init_latents, noise, timestep)
latents = init_latents
return latents
2.stable_diffusion_reference-only
ref_image---->ref_image_latent(vae_encode)–>ref_xt(+noise)–>参考图像与unet注意力层的连接(基于cross_att,这块代码没太看懂)
with self.progress_bar(total=num_inference_steps) as progress_bar:
for i, t in enumerate(timesteps):
# expand the latents if we are doing classifier free guidance
latent_model_input = torch.cat([latents] * 2) if do_classifier_free_guidance else latents
latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)
# ref only part
noise = randn_tensor(
ref_image_latents.shape, generator=generator, device=device, dtype=ref_image_latents.dtype
)
ref_xt = self.scheduler.add_noise(
ref_image_latents,
noise,
t.reshape(
1,
),
)
ref_xt = torch.cat([ref_xt] * 2) if do_classifier_free_guidance else ref_xt
ref_xt = self.scheduler.scale_model_input(ref_xt, t)
MODE = "write"
self.unet(
ref_xt,
t,
encoder_hidden_states=prompt_embeds,
cross_attention_kwargs=cross_attention_kwargs,
return_dict=False,
)
# predict the noise residual
MODE = "read"
noise_pred = self.unet(
latent_model_input,
t,
encoder_hidden_states=prompt_embeds,
cross_attention_kwargs=cross_attention_kwargs,
return_dict=False,
)[0]
# perform guidance
if do_classifier_free_guidance:
noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)
if do_classifier_free_guidance and guidance_rescale > 0.0:
# Based on 3.4. in https://arxiv.org/pdf/2305.08891.pdf
noise_pred = rescale_noise_cfg(noise_pred, noise_pred_text, guidance_rescale=guidance_rescale)
# compute the previous noisy sample x_t -> x_t-1
latents = self.scheduler.step(noise_pred, t, latents, **extra_step_kwargs, return_dict=False)[0]
# call the callback, if provided
if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
progress_bar.update()
if callback is not None and i % callback_steps == 0:
step_idx = i // getattr(self.scheduler, "order", 1)
callback(step_idx, t, latents)