mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-12-27 21:11:59 +00:00
fb3c9be1fd
The term "artifacts" better represents the purpose of this API, which handles outputs generated by API executions, eventually stored objects that can be of served by any storage interface (file, objects). This aligns better with the industry convention of 'artifacts' (build outputs, process results) rather than generic 'files'. 'files' would be appropriate if the goal was to store and retrieve files purely. Additionally, in our context, artifact is a better term since it will handle: * Data produced by SDG (Synthetic Data Generation) - as input * Output of a trained model - as output Signed-off-by: Sébastien Han <seb@redhat.com> |
||
|---|---|---|
| .. | ||
| agents | ||
| artifacts | ||
| batch_inference | ||
| benchmarks | ||
| common | ||
| datasetio | ||
| datasets | ||
| eval | ||
| inference | ||
| inspect | ||
| models | ||
| post_training | ||
| providers | ||
| safety | ||
| scoring | ||
| scoring_functions | ||
| shields | ||
| synthetic_data_generation | ||
| telemetry | ||
| tools | ||
| vector_dbs | ||
| vector_io | ||
| __init__.py | ||
| datatypes.py | ||
| resource.py | ||
| version.py | ||