20260706:新增一些文章
This commit is contained in:
43
concepts/mlp-width-tapering.md
Normal file
43
concepts/mlp-width-tapering.md
Normal file
@@ -0,0 +1,43 @@
|
||||
---
|
||||
title: "MLP Width Tapering(MLP 宽度渐缩)"
|
||||
created: 2026-06-29
|
||||
updated: 2026-06-29
|
||||
type: concept
|
||||
tags: [language-model, mlp, architecture, efficiency, transformer]
|
||||
sources: [[tapered-language-models]]
|
||||
confidence: high
|
||||
---
|
||||
|
||||
# MLP Width Tapering
|
||||
|
||||
> [[depth-aware-capacity-allocation|深度感知容量分配]] 的具体实例化:在固定总参数量下,将 Transformer/LLM 中 MLP 的中间层宽度(d_ff)沿深度单调递减。
|
||||
|
||||
## 为什么选 MLP
|
||||
|
||||
| 组件 | 参数占比 | 跨架构通用性 | 调节轴 |
|
||||
|------|---------|-------------|--------|
|
||||
| **MLP (FFN)** | 主导(~2/3) | 所有架构共享 | 单一标量 d_ff |
|
||||
| Attention | 次之 | 架构差异大 | 多头/维度耦合 |
|
||||
| LayerNorm | 极小 | 通用 | 无意义 |
|
||||
|
||||
MLP 是自然的 taper 目标:参数量最大、架构通用、调节维度单一干净。
|
||||
|
||||
## Cosine Schedule
|
||||
|
||||
推荐的衰减方式见 [[cosine-taper-schedule|余弦衰减调度]]:
|
||||
|
||||
> w_ℓ = w_max · (cos(πℓ/(2L)))^p
|
||||
|
||||
- ℓ ∈ [0, L-1]:层索引
|
||||
- p:控制陡峭度
|
||||
- 最优范围:1.50× → 0.50× baseline d_ff
|
||||
|
||||
## 效果
|
||||
|
||||
440M Transformer:uniform 16.28 → cosine taper 14.44 perplexity(零额外参数/计算)。
|
||||
|
||||
## 相关概念
|
||||
|
||||
- [[depth-aware-capacity-allocation|深度感知容量分配]]
|
||||
- [[cosine-taper-schedule|余弦衰减调度]]
|
||||
- [[tapered-language-models|Tapered Language Models]]
|
||||
Reference in New Issue
Block a user