paper-with-me

홈 › Papers

Merging Text Transformer Models from Different Initializations

2024-03-01 · Neha Verma, Maha Elbayad

Recent work on permutation-based model merging has shown impressive low- or zero-barrier mode connectivity between models from completely different initializations. However, this line of work has not yet extended to the Transformer architecture, despite its dominant popularity in the language domain. Therefore, in this work, we investigate the extent to which separate Transformer minima learn similar features, and propose a model merging technique to investigate the relationship between these minima in the loss landscape. The specifics of the architecture, like its residual connections, multi-headed attention, and discrete, sequential input, require specific interventions in order to compute model permutations that remain within the same functional equivalence class. In merging these models with our method, we consistently find lower loss barriers between minima compared to model averaging, across models trained on a masked-language modeling task or fine-tuned on a language understanding benchmark. Our results show that the minima of these models are less sharp and isolated than previously understood, and provide a basis for future work on merging separately trained Transformer models.

📄 PDF Abstract BibTeX arXiv:2403.00986

Code (1)

nverma1/merging-text-transformers 공식 구현 pytorch

Tasks

Language ModelingLanguage ModellingMasked Language Modeling

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
BPE Byte Pair Encoding, or BPE, is a subword segmentation algorithm that encodes rare and unknown words as sequences of subword units. The intuition is that various word…
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…
Multi-Head Attention 설명 없음
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Dense Connections Dense Connections, or Fully Connected Connections, are a type of layer in a deep neural network that use a linear operation where every input is connected to every output…

Similar Papers 제목 키워드 기반

Foldable SuperNets: Scalable Merging of Transformers with Different Initializations and Tasks

2024-10-02 · Edan Kinderman, Itay Hubara, Haggai Maron, Daniel Soudry

Many recent methods aim to merge neural networks (NNs) with identical architectures trained on different tasks to obtain a single multi-task model. Most existing works tackle the simpler setup of merging NNs initialized …

Knowledge Distillation

Merging Decision Transformers: Weight Averaging for Forming Multi-Task Policies

2023-03-14 · Daniel Lawson, Ahmed H. Qureshi

Recent work has shown the promise of creating generalist, transformer-based, models for language, vision, and sequential decision-making problems. To create such models, we generally require centralized training objectiv…

Decision MakingMuJoCoSequential Decision Making

Playing the lottery with rewards and multiple languages: lottery tickets in RL and NLP

2019-06-06 · ICLR 2020 1 · Haonan Yu, Sergey Edunov, Yuandong Tian, Ari S. Morcos

The lottery ticket hypothesis proposes that over-parameterization of deep neural networks (DNNs) aids training by increasing the probability of a "lucky" sub-network initialization being present rather than by helping th…

image-classificationImage ClassificationReinforcement LearningReinforcement Learning (RL)

Vanishing Feature: Diagnosing Model Merging and Beyond

2024-02-05 · Xingyu Qu, Samuel Horvath

Model merging offers an efficient way to combine pre-trained neural networks but often suffers from inconsistent performance, especially when merging models with different initializations. We identify the ``vanishing fea…

Linear Mode ConnectivitymodelRe-basin

What’s Hidden in a One-layer Randomly Weighted Transformer?

2021-11-01 · EMNLP 2021 11 · Sheng Shen, Zhewei Yao, Douwe Kiela, Kurt Keutzer 외

We demonstrate that, hidden within one-layer randomly weighted neural networks, there exist subnetworks that can achieve impressive performance, without ever modifying the weight initializations, on machine translation t…

Machine TranslationTranslation