feat(python): Embedding model tuner #1221
Open
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
AyushExel
commented
Apr 16, 2024
| pa.Table.from_pydict(relevant_docs), | ||
| save_dir / "relevant_docs.lance", | ||
| mode=mode, | ||
| ) |
There was a problem hiding this comment.
TODO:
- Put all 3 parts of the dataset into a single lance table
Probably doable but the current design treats docs as a single unit with any number of query and response pairs possible..
AyushExel
commented
Apr 16, 2024
| f"model = get_registry().get('sentence-transformers').create(name='./{self.path}')" # noqa | ||
| ) | ||
|
|
||
| def _wandb_callback(self, score, epoch, steps): |
There was a problem hiding this comment.
Future TODO:
- This integration doesn't work. Investigate
AyushExel
commented
Apr 16, 2024
| res = model.evaluate(ds) | ||
| assert res is not None | ||
|
|
||
|
|
There was a problem hiding this comment.
Future TODO:
- This is an umbrella test. Need more granular ones
AyushExel
commented
Apr 22, 2024
| model: Any, | ||
| trainset: QADataset, | ||
| valset: Optional[QADataset] = None, | ||
| path: Optional[str] = "~/.lancedb/embeddings/models", |
There was a problem hiding this comment.
TODO: check conflict with .lance files
AyushExel
commented
Apr 22, 2024
| llm: BaseLLM, | ||
| qa_generate_prompt_tmpl: str = DEFAULT_PROMPT_TMPL, | ||
| num_questions_per_chunk: int = 2, | ||
| ) -> "QADataset": |
There was a problem hiding this comment.
Accept generators
AyushExel
commented
Apr 22, 2024
| nodes: List["TextChunk"], | ||
| queries: Dict[str, str], | ||
| relevant_docs: Dict[str, List[str]], | ||
| ) -> "QADataset": |
There was a problem hiding this comment.
TODO: test api with WANDS dataset
AyushExel
commented
Apr 22, 2024
| from abc import ABC, abstractmethod | ||
|
|
||
|
|
||
| class BaseEmbeddingTuner(ABC): |
There was a problem hiding this comment.
TODO: get rid of this
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.
Solves #1021
Design and usage doc - https://www.notion.so/LanceDB-High-Level-Specs-From-ML-perspective-f9b7470b1e4e4c9e8371ad28b574c185?pvs=4#d6a4f29edf3d4ced9954ab8a913ef9f0
Benchmarks for 5 epochs on 65-35 train test split - https://wandb.ai/cayush/lancedb_finetune?nw=nwusercayush