alibaba / alibaba/x-deeplearning

TDM 旧模型 新增训练数据(有新的item加入)时,应该如何初始化树文件

Open
#347 2 comments 0 reactions 0 assignees View on GitHub
Dominant language
PureBasic
Stars
4.3k
Forks
1k
PR merge metrics
No merged PRs in 30d

Description

有个问题想请教
现在我已经有了0-15天训练数据训练出的模型,我想在上次训练的基础上在训练16-30天的数据(会有不少新的item加入),这时userbehavoir_tree.pb 和 leaf.id 文件中只有老的item id 节点,我应该如何加入新的item呀,
如果使用tree_init.py ,我理解之前的模型就会失效 因为node节点变化太大了 node_emb就得完全重新学
如果使用tree_cluster.py, 新加入的item又没有embedding用来聚类

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reading tree_init.py and tree_cluster.py, then inspect how userbehavoir_tree.pb, leaf.id, and node_emb are produced and consumed. A complete result would define and validate an agreed approach for adding new item IDs during incremental training without discarding the existing model.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.