This is Unstract's python package which helps to configure to a number of different LLMs, Embeddings and VectorDBs.
The following LLMs are supported:
| LLM | Version |
|---|---|
| OpenAI | 1.3.9 |
| Azure OpenAI | 1.3.9 |
| Anthropic | 0.7.8 |
| PaLM | 0.3.1 |
| Replicate | 0.22.0 |
| AnyScale | 0.5.165 |
| Mistral | 0.0.8 |
The following Embeddings are supported:
| Embedding | Version |
|---|---|
| OpenAI | 1.3.9 |
| Azure OpenAI | 1.3.9 |
| Qdrant FastEmbed | 0.1.3 |
| HuggingFace | 0.0.1 |
| PaLM | 0.3.1 |
The following VectorDBs are supported:
| Vector DB | Version |
|---|---|
| Milvus | 2.3.4 |
| Pinecone | 2.2.4 |
| Postgres | 0.2.4 |
| Qdrant | 1.7.0 |
| Supabase | 2.2.1 |
| Weaviate | 3.25.3 |
To get started with local development,
- Create and source a virtual environment if you haven't already following these steps.
- If you're using Mac, install the below library needed for PyMSSQL
brew install pkg-config freetds
- Install the required dependencies with
pdm install