Lightning-AI/lit-llama

Resume in the pretraining code

Open

#359 建立於 2023年6月2日

在 GitHub 查看
 (2 留言) (1 反應) (0 負責人)Python (5,533 star) (473 fork)batch import
help wanted

描述

I would like to request a new feature in the code: the ability to resume training from a checkpoint.

Currently, the code can save a checkpoint of the model's state at any point during training. However, there is no way to resume training from a checkpoint.

The code can save two things along with the model state_dict: 1)the optimizer, 2)the id of the last example it has seen (assuming the data is fed sequentially to the model not randomly)

貢獻者指南

技術棧
pythonpytorch
領域
machine learningai
議題類型
feature
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
1-2 days
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
fresh
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
clear
前置要求
Basic Python and PyTorch knowledgeUnderstanding of training loop checkpoints
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
60
研究方向
First, examine the current checkpoint saving logic in the pretraining code (likely in the `train.py` or similar file). The implementation should add a ` resume from checkpoint` argument that loads the model state dict, optimizer state dict, and the last processed example ID. Then, modify the data loading to skip already processed examples and continue training from that point. Refer to the existing saving logic and any comments in the code for guidance.