facebookresearch/ParlAI

Support tied positional weights in transformers

Open

#2,184 创建于 2019年11月19日

在 GitHub 查看
 (1 评论) (0 反应) (0 负责人)Python (10,371 star) (2,089 fork)batch import
EnhancementH1 2020Help WantedMediumdonotreap

描述

Add a flag to use tied positional embeddings in transformer/generator and transformer/retrieval, and implement the tied weights. Should be False by default for backwards compatibility, and upgrade_opt should be used to migrate old models.

贡献者指南

技术栈
python
领域
machine learning
议题类型
feature
难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。
3
预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。
1-2 days
活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。
fresh
清晰度议题是否清楚说明期望改动、验收标准和下一步。
clear
前置要求
Understanding of transformer architecturesFamiliarity with ParlAI codebasePython programming
新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。
30
研究方向
Examine the transformer generator and retrieval modules in ParlAI (likely under parlai/agents/transformer/). Review how positional embeddings are currently implemented. Add a new boolean flag for tied positional weights, defaulting to False. Implement the tied weights logic and use upgrade opt to handle migration of old models. Check issue #2184 comments for any additional guidance.
Support tied positional weights in transformers · facebookresearch/ParlAI#2184 | Good First Issue