Skip to content

fix(bria_fibo): fix guidance_embeds, prompt_embeds, tensor-image and multi-image crashes #13981

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
akshan-main wants to merge 1 commit into
base: main
Choose a base branch
from
+23 −11
Open

fix(bria_fibo): fix guidance_embeds, prompt_embeds, tensor-image and multi-image crashes #13981

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions src/diffusers/models/transformers/transformer_bria_fibo.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -470,7 +470,7 @@ def __init__(
self.time_embed = BriaFiboTimestepProjEmbeddings(embedding_dim=self.inner_dim, time_theta=time_theta)

if guidance_embeds:
self.guidance_embed = BriaFiboTimestepProjEmbeddings(embedding_dim=self.inner_dim)
self.guidance_embed = BriaFiboTimestepProjEmbeddings(embedding_dim=self.inner_dim, time_theta=time_theta)

self.context_embedder = nn.Linear(self.config.joint_attention_dim, self.inner_dim)
self.x_embedder = torch.nn.Linear(self.config.in_channels, self.inner_dim)
Expand Down Expand Up @@ -555,7 +555,7 @@ def forward(

temb = self.time_embed(timestep, dtype=hidden_states.dtype)

if guidance:
if guidance is not None:
temb += self.guidance_embed(guidance, dtype=hidden_states.dtype)

encoder_hidden_states = self.context_embedder(encoder_hidden_states)
Expand Down
14 changes: 10 additions & 4 deletions src/diffusers/pipelines/bria_fibo/pipeline_bria_fibo.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -260,6 +260,11 @@ def encode_prompt(
)
prompt_embeds = prompt_embeds.to(dtype=self.transformer.dtype)
prompt_layers = [tensor.to(dtype=self.transformer.dtype) for tensor in prompt_layers]
else:
raise ValueError(
"`prompt_embeds` cannot be passed on its own; this pipeline also needs the per-layer embeddings "
"computed from `prompt`. Please pass `prompt` instead."
)

if guidance_scale > 1:
if isinstance(negative_prompt, list) and negative_prompt[0] is None:
Expand Down Expand Up @@ -773,10 +778,11 @@ def __call__(
for scaled_latent in latents_scaled:
curr_image = self.vae.decode(scaled_latent.unsqueeze(0), return_dict=False)[0]
curr_image = self.image_processor.postprocess(curr_image.squeeze(dim=2), output_type=output_type)
image.append(curr_image)
if len(image) == 1:
image = image[0]
else:
if output_type == "np":
image.append(curr_image[0])
else:
image.extend(curr_image)
if output_type == "np":
image = np.stack(image, axis=0)

# Offload all models
Expand Down
16 changes: 11 additions & 5 deletions src/diffusers/pipelines/bria_fibo/pipeline_bria_fibo_edit.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -418,6 +418,11 @@ def encode_prompt(
)
prompt_embeds = prompt_embeds.to(dtype=self.transformer.dtype)
prompt_layers = [tensor.to(dtype=self.transformer.dtype) for tensor in prompt_layers]
else:
raise ValueError(
"`prompt_embeds` cannot be passed on its own; this pipeline also needs the per-layer embeddings "
"computed from `prompt`. Please pass `prompt` instead."
)

if guidance_scale > 1:
if isinstance(negative_prompt, list) and negative_prompt[0] is None:
Expand Down Expand Up @@ -807,7 +812,7 @@ def __call__(
prompt_layers = prompt_layers + [prompt_layers[-1]] * (total_num_layers_transformer - len(prompt_layers))

# Preprocess image
if image is not None and not (isinstance(image, torch.Tensor) and image.size(1) == self.latent_channels):
if image is not None:
image = self.image_processor.resize(image, height, width)
image = self.image_processor.preprocess(image, height, width)

Expand Down Expand Up @@ -996,10 +1001,11 @@ def __call__(
for scaled_latent in latents_scaled:
curr_image = self.vae.decode(scaled_latent.unsqueeze(0), return_dict=False)[0]
curr_image = self.image_processor.postprocess(curr_image.squeeze(dim=2), output_type=output_type)
image.append(curr_image)
if len(image) == 1:
image = image[0]
else:
if output_type == "np":
image.append(curr_image[0])
else:
image.extend(curr_image)
if output_type == "np":
image = np.stack(image, axis=0)

# Offload all models
Expand Down
Loading