forked from phoenix-oss/llama-stack-mirror
There is a torchtune release this morning https://github.com/pytorch/torchtune/releases/tag/v0.5.0 and breaks post training apis ## test spinning up server and the post training works again after the fix <img width="1314" alt="Screenshot 2024-12-20 at 4 08 54 PM" src="https://github.com/user-attachments/assets/dfae724d-ebf0-4846-9715-096efa060cee" /> ## Note We need to think hard of how to avoid this happen again and have a fast follow up on this after holidays |
||
|---|---|---|
| .. | ||
| agents | ||
| datasetio | ||
| eval | ||
| inference | ||
| ios/inference | ||
| memory | ||
| post_training/torchtune | ||
| safety | ||
| scoring | ||
| telemetry | ||
| tool_runtime/brave_search | ||
| __init__.py | ||