fix dtype bugs #72
Merged
fix dtype bugs #72
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
There was a problem hiding this comment.
Pull Request Overview
This PR addresses dtype-related issues and optimizes device memory management by updating model initialization and data transfers across multiple modules. Key changes include:
- Addition of a dtype parameter when instantiating FluxImagePipeline objects in tools modules.
- Replacement of torch.cuda.empty_cache() with a unified empty_cache() utility for cross-platform cache clearing.
- Updates to model and tensor device conversion in pipelines and VAE modules to ensure proper dtype propagation.
Reviewed Changes
Copilot reviewed 9 out of 9 changed files in this pull request and generated 2 comments.
Show a summary per file
| File | Description |
|---|---|
| diffsynth_engine/utils/platform.py | Introduces a new cross-platform cache clearing utility. |
| diffsynth_engine/tools/flux_replace_tool.py | Updates FluxImagePipeline initialization by adding a dtype parameter. |
| diffsynth_engine/tools/flux_reference_tool.py | Similar update as flux_replace_tool to include dtype in model instantiation. |
| diffsynth_engine/tools/flux_outpainting_tool.py | Modifies pipeline initialization by adding dtype support. |
| diffsynth_engine/tools/flux_inpainting_tool.py | Adjusts pipeline initialization to pass dtype correctly. |
| diffsynth_engine/pipelines/flux_image.py | Improves device model management and cache clearing during noise prediction and latent preparation. |
| diffsynth_engine/pipelines/base.py | Refactors device transfers to utilize empty_cache() and updates tensor conversions using model dtype. |
| diffsynth_engine/models/vae/vae.py | Stores device and dtype on the VAE models to standardize downstream tensor conversions. |
| diffsynth_engine/models/utils.py | Minor cleanup of code comments. |
| return noise | ||
|
|
||
| def encode_image(self, image: torch.Tensor) -> torch.Tensor: | ||
| image = image.to(self.device, self.vae_encoder.dtype) |
There was a problem hiding this comment.
[nitpick] Consider using keyword arguments for the to() method (e.g., image.to(device=self.device, dtype=self.vae_encoder.dtype)) to improve code clarity.
Suggested change
| image = image.to(self.device, self.vae_encoder.dtype) | |
| image = image.to(device=self.device, dtype=self.vae_encoder.dtype) |
sir1st-inc
approved these changes
May 23, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.