marqo
marqo copied to clipboard
marqo-ai
Unified embedding generation and search engine. Also available on cloud - cloud.marqo.ai
**Describe the solution you'd like** expose the size of an index via the index stats endpoint
Give warning/error when almost or completely out of disk. Errors if out of disk: - On startup - On attempting to load a new model Log warnings: - When close...
**Describe the bug** A nice error to reject injecting multimodal fields and string fields into the same field ``` doc1 = {"image":"https://image.png/"} doc2 = {"image": {"img": "https://some-thing/", "title": "dog walker"...
Have a guide we can point users to, to make their large data ingestion experience better. Suggestions would include: - GPU instances - No auto refreshing - Using non-tensorable fields
Log torch info on startup. This helps debugging. This information will help: ``` In [1]: import torch to In [2]: torch.__version__ Out[2]: '2.0.0.dev20230118+cu116' In [3]: torch.version.cuda Out[3]: '11.6' ```
**Describe the solution you'd like** When throttling is enabled and marqo tries to connect to redis while redis is down, a warning is logged. It should also do 1 or...
**Is your feature request related to a problem? Please describe.** For use cases where highlights and precision take less priority than economically storing data, we want to store a single...
**Describe the bug** Number of shards are defined at index creation time and aren't updated **Suggested fix** Get the updated shard count when get index settings is requested
**Is your feature request related to a problem? Please describe.** Please provide a solution for backups and restores at an index level. **Describe the solution you'd like** I need to...
## Feature Limit the number of tensor fields that can be included in the searchable attributes, configurable by environment variables. ## Changes - Environment variable `MARQO_MAX_SEARCHABLE_TENSOR_ATTRIBUTES` sets the maximum number...
