talemate/docs/chromadb.md

## ChromaDB

If you want chromaDB to use the more accurate (but much slower) instructor embeddings add the following to `config.yaml`:

```yaml
chromadb:
    embeddings: instructor
    instructor_device: cpu
    instructor_model: hkunlp/instructor-xl"
```

You will need to restart the backend for this change to take effect.

Note that the first time you do this it will need to download the instructor model you selected. This may take a while, and the talemate backend will be un-responsive during that time.

Once the download is finished, if talemate is still un-responsive, try reloading the front-end to reconnect. When all fails just restart the backend as well.

### GPU support

If you want to use the instructor embeddings with GPU support, you will need to install pytorch with CUDA support. 

To do this on windows, run `install-pytorch-cuda.bat` from the project root. Then change your device in the config to `cuda`:

```yaml
chromadb:
    embeddings: instructor
    instructor_device: cuda
    instructor_model: hkunlp/instructor-xl"
```

Instructor embedding models:

- `hkunlp/instructor-base` (smallest / fastest)
- `hkunlp/instructor-large` 
- `hkunlp/instructor-xl` (largest / slowest) - requires about 5GB of GPU memory