642 train 3d model with lucchi data #650
Draft
Conversation
This file contains 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
* Update faq.md * Add described answer to the issue
Updates to SAM 3d training --------- Co-authored-by: Constantin Pape <[email protected]>
…er (#645) Add simple 3d wrapper and enable freezing the encoder in sam 3d wrapper, simplify lora support
Minor fix to trainable sam model functionality
Clean up interfaces related to 3d models and PEFT
| assert os.path.exists(image_path), image_path | ||
|
|
||
| # Perform segmentation only on the semantic class | ||
| for i, (semantic_class_name, _) in enumerate(semantic_class_map.items()): |
There was a problem hiding this comment.
I would suggest to remove this part already, it doesn't make sense in the context here.
| def predict(args): | ||
|
|
||
| device = "cuda" if torch.cuda.is_available() else "cpu" | ||
| if args.checkpoint_path is not None: |
There was a problem hiding this comment.
I am not sure why you would ever run prediction without a checkpoint. I would not make this optional.
| lora_rank=4, | ||
| model_type=args.model_type, | ||
| checkpoint_path=cp_path | ||
| ) |
There was a problem hiding this comment.
This will not work to actually load the checkpoint. Please read the code I send you carefully and see how I use torch_em.util.load_model to get the checkpoint. Not that this quite complex way is now necessary because of the way how we mix the 3d adapter and LoRA. I will think about a better way at some point, but for now we need to use this work-around.
| return raw | ||
|
|
||
|
|
||
| class LucchiSegmentationDataset(SegmentationDataset): |
There was a problem hiding this comment.
This can now be removed.
… default_sam_dataset
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.