PG-SFT: Balancing Capability Acquisition and Retention in Offline Agent Fine-Tuning
Ronghua Li, Zi Liang, Zhishan Li, Shinan Liu
実装難易度
Hard
推論・学習コスト
High
想定用途
生成
概要
Supervised fine-tuning (SFT) on offline agent trajectories is the standard approach for training specialized tool-using agents, but forcing models to imitate reasoning and actions token by token may harm other capabilities (e.g., general reasoning, tool calling, code generation) of the base model. In this work, we focus on studying \emph{how to better balance the trade-off between acquiring new…
何が新しいか
Supervised fine-tuning (SFT) on offline agent trajectories is the standard approach for training specialized tool-using agents, but forcing models to imitate reasoning and actions token by token may…
何に使えるか
生成
実装情報
- Paper URL
- あり
実装チェックリスト
実装または配布ページ
要確認Paper onlyの可能性があるため再実装前提で確認してください。
一次情報リンク
OKPaper
検証しやすさ
要確認公式実装が見つからないため、論文から再実装する前提です。
計算資源
要確認学習や高解像度推論ではGPUメモリと実行時間に注意が必要です。
ライセンス
未取得配布元のLICENSE、モデルカード、Paperの利用条件を確認してください。
商用利用
未取得研究利用限定、データセット由来制限、API規約の有無を確認してください。
自社データで試すなら
製造業・材料開発のExcel/CSVデータに落とし込むための最初の手順です。
- 1まず自社データを、入力条件、目的変数、評価したい指標に分けて整理します。
- 2LightGBMやRandom Forestなどのベースラインを先に作り、この手法と比較します。
- 3評価指標はR2/RMSE、AUC、異常検知の再現率、実験回数削減率など、現場の意思決定に近いものを選びます。
- 4SHAPや特徴量重要度で、効いている因子が物理・化学・工程知識と矛盾しないか確認します。
実装難易度
Hard - 公式実装が見つからないため、論文から再実装する前提です。
必要リソース
- GPU目安: High
- データセット: 論文・リポジトリ側の指定を確認してください。
- 学習要否: 再学習や評価環境の準備が必要になる可能性があります。
- 学習や高解像度推論ではGPUメモリと実行時間に注意が必要です。
実務で使う場合の注意点
- ライセンスと商用利用条件は、Paper / GitHub / Hugging Face の配布元で確認してください。
- 精度、再現性、計算コストはデータセットや評価条件に依存します。
- 個人情報や機密データを扱う場合は、入力データの保存先と外部API利用条件を確認してください。
関連記事
A Matched-Budget Audit Framework for Recaptioned Image-Text Supervision Distributions
Recaptioned image-text corpora are now standard for text-to-image (T2I) training, with vision--language model
Know When to Hold 'em: Correct-Token Retention in Uniform-State Diffusion Language Models
Uniform-state diffusion models (USDMs) can revise any token at any denoising step, which lets them correct the
AutoGUIWorld: Image Generators as Visual World Models for GUI Agent
GUI agents require high-quality interaction trajectories to learn how software environments respond to actions
Gumbel Straight Flow: Distilling Autoregressive Models into One-step Flow Maps
We present Gumbel Straight Flow (GSF), a continuous flow map language model that leverages the noise-data coup