Add support for finetune guard classifier #325
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
Signed-off-by: Vibhu Jawa <[email protected]>
Signed-off-by: Vibhu Jawa <[email protected]>
VibhuJawa
changed the title
Signed-off-by: Vibhu Jawa <[email protected]>
Signed-off-by: Vibhu Jawa <[email protected]>
VibhuJawa
commented
Nov 18, 2024
| - Demonstrated strong zero-shot detection capabilities on novel attacks | ||
| - Particularly effective at identifying trigger patterns in partially poisoned datasets | ||
|
|
||
| Dataset Format: |
There was a problem hiding this comment.
TODO: Emphasize more about english
sarahyurick
reviewed
Nov 21, 2024
There was a problem hiding this comment.
Added a few typo fixes but looks good so far!
When you get a chance can you add it to the examples/ and nemo_curator/scripts/ folders? And to the documentation? You should be able to reference #361 which has all the files that should be created/updated.
| class AegisModel(nn.Module): | ||
| def __init__( | ||
| self, | ||
| pretrained_model_name_or_path: str, | ||
| peft_model_name_or_path: str, | ||
| dtype: torch.dtype, | ||
| token: str, | ||
| token: Optional[Union[str, bool]], | ||
| add_fintune_gaurd: bool = False, |
There was a problem hiding this comment.
Suggested change
| add_fintune_gaurd: bool = False, | |
| add_finetune_guard: bool = False, |
| ): | ||
| super().__init__() | ||
| base_model = AutoModelForCausalLM.from_pretrained( | ||
| pretrained_model_name_or_path, torch_dtype=dtype, token=token | ||
| ) | ||
| self.model = PeftModel.from_pretrained(base_model, peft_model_name_or_path) | ||
| self.autocast = autocast | ||
| self.add_fintune_gaurd = add_fintune_gaurd |
There was a problem hiding this comment.
Suggested change
| self.add_fintune_gaurd = add_fintune_gaurd | |
| self.add_finetune_guard = add_finetune_guard |
| ): | ||
| super().__init__() | ||
| base_model = AutoModelForCausalLM.from_pretrained( | ||
| pretrained_model_name_or_path, torch_dtype=dtype, token=token | ||
| ) | ||
| self.model = PeftModel.from_pretrained(base_model, peft_model_name_or_path) | ||
| self.autocast = autocast | ||
| self.add_fintune_gaurd = add_fintune_gaurd | ||
| if self.add_fintune_gaurd: |
There was a problem hiding this comment.
Suggested change
| if self.add_fintune_gaurd: | |
| if self.add_finetune_guard: |
| pad_token_id=0, | ||
| ) | ||
| def _forward(self, batch): | ||
| if self.add_fintune_gaurd: |
There was a problem hiding this comment.
Suggested change
| if self.add_fintune_gaurd: | |
| if self.add_finetune_guard: |
| if self.config.add_finetune_guard: | ||
| if self.config.finetune_guard_path is None: | ||
| raise ValueError( | ||
| "finetune_guard_path must be provided if add_fine_guard is True" |
There was a problem hiding this comment.
Suggested change
| "finetune_guard_path must be provided if add_fine_guard is True" | |
| "finetune_guard_path must be provided if add_finetune_guard is True" |
| peft_model_name_or_path=self.config.peft_model_name_or_path, | ||
| dtype=self.config.dtype, | ||
| token=self.config.token, | ||
| add_fintune_gaurd=self.config.add_finetune_guard, |
There was a problem hiding this comment.
Suggested change
| add_fintune_gaurd=self.config.add_finetune_guard, | |
| add_finetune_guard=self.config.add_finetune_guard, |
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.
This PR adds support for FineTuneGuard Model