Originally created by @ard0gg on GitHub (Oct 10, 2025).
I'm having issue using a the nomic-embed-text model with ollama. I've updated the embedding dimensions to 768 per nomic documentation. However, whenever I try to ingest data I get this error
minne_app | 2025-10-10T17:39:59.331187Z ERROR main: Worker process error: Ingestion Processing error: Database error: Incorrect vector dimension (768). Expected a vector of 1536 dimension.
Originally created by @ard0gg on GitHub (Oct 10, 2025).
I'm having issue using a the nomic-embed-text model with ollama. I've updated the embedding dimensions to 768 per nomic [documentation](https://docs.nomic.ai/platform/embeddings-and-retrieval/text-embedding). However, whenever I try to ingest data I get this error
```bash
minne_app | 2025-10-10T17:39:59.331187Z ERROR main: Worker process error: Ingestion Processing error: Database error: Incorrect vector dimension (768). Expected a vector of 1536 dimension.
```
Thanks for reporting! The only current workaround would be to ingest content with the default embedding dimensions, and then switch. I’ll get to fixing this at the end of the weekend!
@perstarkse commented on GitHub (Oct 10, 2025):
Thanks for reporting! The only current workaround would be to ingest content with the default embedding dimensions, and then switch. I’ll get to fixing this at the end of the weekend!
I've now implemented a fix in commit b7344644dc, and it should be in the latest release
@perstarkse commented on GitHub (Oct 12, 2025):
I've now implemented a fix in commit b7344644dc2daa3c4ab5eb680e8a626ee4dfc526, and it should be in the latest release
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @ard0gg on GitHub (Oct 10, 2025).
I'm having issue using a the nomic-embed-text model with ollama. I've updated the embedding dimensions to 768 per nomic documentation. However, whenever I try to ingest data I get this error
@perstarkse commented on GitHub (Oct 10, 2025):
Thanks for reporting! The only current workaround would be to ingest content with the default embedding dimensions, and then switch. I’ll get to fixing this at the end of the weekend!
@perstarkse commented on GitHub (Oct 12, 2025):
I've now implemented a fix in commit
b7344644dc, and it should be in the latest release@ard0gg commented on GitHub (Oct 12, 2025):
Thanks for the quick fix!