PyPI page
Home page
Author:
None
Summary:
PyTorch-based trainer for Agent trajectory datasets — SFT, DPO, GRPO
Latest version:
0.1.1
Required dependencies:
accelerate
|
click
|
pydantic
|
pyyaml
|
torch
|
tqdm
|
transformers
Optional dependencies:
mcp
|
peft
|
pytest
|
ruff
|
wandb
Downloads last day:
0
Downloads last week:
2
Downloads last month:
30