superduper
superduper copied to clipboard
superduper-io
Superduper: Integrate AI models and machine learning workflows with your database to implement custom AI applications, without moving your data. Including streaming inference, scalable model hosting,...
For the built-in Leaf object in superduperdb, reduce the amount of information through special references For example **Now** ```python from superduperdb.components.datatype import pickle_serializer from superduperdb import Document Document({'id': 123, 'x':...
We create an S3 DataType with a parameter, pre_download. ## The logic when pre_download is True and encodable as File **During encoding:** Download the file from S3 and create a...
Currently if we delete a component it deletes all artifacts. This is a problem if these are shared between components - e.g. templates and applications.
I 'm trying to run superduper without any configuration, but it fails: ``` 2024-May-17 00:16:19.96| ERROR | 8caba5037a98| superduperdb.base.build:72 | Error initializing to DataBackend Client: localhost:27017: [Errno 111] Connection refused...
### Contact Details [Optional] _No response_ ### System Information main ### What happened? ```python from superduperdb import ObjectModel from superduperdb.components.graph import Graph, document_node, input_node def func1(x, k=100): return x +...
The following table is printed only after the data layer has been built. This is nice for info, but hard for debugging, because if the data layer fails, the table...
it does not seem to work (I removed the fields as this is public channel) ``` | 2024-May-29 20:00:39.45| INFO | statefulset-0| superduperdb.base.build:53 | Connecting to Metadata Client with URI:...
Currently, the listener uses UUID as predict_id, making it difficult to find model results in the database since it’s a string of meaningless characters. The predict_id should have the following...
```python def dict(self) -> 'Document': """A dictionary representation of the component.""" from superduperdb import Document from superduperdb.components.datatype import Artifact, File r = super().dict() s = self.artifact_schema for k in s.fields:...
`test/smoke/test_vector_search.py` depends on `database_with_default_encoders_and_model` which depends on `add_models_encoders` which depends on `torch`. However, since `torch` is not installed by default, the `pytest -s test/smoke/test_vector_search.py` returns `NameError: name 'tensor' is not...
Metadata
Owner
Metadata
Superduper: Integrate AI models and machine learning workflows with your database to implement custom AI applications, without moving your data. Including streaming inference, scalable model hosting,...