Neural Question Answering & Semantic Search at Scale. Use modern transformer based models like BERT
Neural Question Answering & Semantic Search at Scale. Use modern transformer based models like BERT to find answers in large document collections
deepset-ai/haystack repo was created 2 years ago and was last updated 10 hours ago.
The project is very popular with an impressive 4642 github stars!
How to Install farm-haystack
You can install farm-haystack using pip
pip install farm-haystack
or add it to a project with poetry
poetry add farm-haystack
- Apache License 2.0
- GitHub Repo
- Scientific/Engineering/Artificial Intelligence
A list of common farm-haystack errors.
Here are some
farm-haystack code examples and snippets.
The farm-haystack package has 195 open issues on GitHub
- Add ADR template for transparent architecture decisions
- OpenSearchDocumentStore: Support cosine similarity on dot_product embedding fields
- allow different filters per query in pipeline evaluation
- Update tutorials, utilities, tests and dependencies to run on Milvus2
- What is the basic machine requirement for deploying ?
- Population based training for fine tuning?
- DPR embedding is not "invalidated" after calling DocumentStore.update_document_meta
- ElasticSearchDocumentStore create_index param: Why is this useful?
- Unable to make connection with existing Elasticstore with data.
- add metadata to summarizer response
- ranker should return scores for later usage
- Autogenerate OpenAPI specs file
- ✨ Add JSON Schema autogeneration for Pipeline YAML files - alternative 2
- Make version of pipeline config/YAML configurable
- Add Haystack CLI utility