KAI AntCore 118M SFT Alpaca
SFT checkpoint for KAI AntCore 118M trained on tatsu-lab/alpaca.
- Base repo:
nfeliperealbr/kai-antcore-118m-v1
- Examples used:
10000
- Epochs:
3
- Sequence length:
512
- Learning rate:
2e-05
- Loss masking: assistant/output tokens only
This is an experimental SFT validation run, not a final aligned assistant.
