TokenFormer: Rethinking Transformer Scaling with Tokenized Model Parameters

• URL: http://arxiv.org/abs/2410.23168v1
• Date: Wed, 30 Oct 2024 16:19:00 GMT
• Title: TokenFormer: Rethinking Transformer Scaling with Tokenized Model Parameters
• Authors: Haiyang Wang, Yue Fan, Muhammad Ferjad Naeem, Yongqin Xian, Jan Eric Lenssen, Liwei Wang, Federico Tombari, Bernt Schiele,
• Abstract summary: We introduce TokenFormer, a scalable architecture for scaling Transformers. By treating model parameters as tokens, we replace all the linear projections in Transformers. Our model scales from 124M to 1.4B parameters by incrementally adding new key-value parameter pairs.
• Score: 102.1116808722299
• License:
• Abstract: Transformers have become the predominant architecture in foundation models due to their excellent performance across various domains. However, the substantial cost of scaling these models remains a significant concern. This problem arises primarily from their dependence on a fixed number of parameters within linear projections. When architectural modifications (e.g., channel dimensions) are introduced, the entire model typically requires retraining from scratch. As model sizes continue growing, this strategy results in increasingly high computational costs and becomes unsustainable. To overcome this problem, we introduce TokenFormer, a natively scalable architecture that leverages the attention mechanism not only for computations among input tokens but also for interactions between tokens and model parameters, thereby enhancing architectural flexibility. By treating model parameters as tokens, we replace all the linear projections in Transformers with our token-parameter attention layer, where input tokens act as queries and model parameters as keys and values. This reformulation allows for progressive and efficient scaling without necessitating retraining from scratch. Our model scales from 124M to 1.4B parameters by incrementally adding new key-value parameter pairs, achieving performance comparable to Transformers trained from scratch while greatly reducing training costs. Code and models are available at \url{https://github.com/Haiyang-W/TokenFormer}.

Related papers

• ConvMixFormer- A Resource-efficient Convolution Mixer for Transformer-based Dynamic Hand Gesture Recognition [5.311735227179715]
  We explore and devise a novel ConvMixFormer architecture for dynamic hand gestures. The proposed method is evaluated on NVidia Dynamic Hand Gesture and Briareo datasets. Our model has achieved state-of-the-art results on single and multimodal inputs.
  arXiv  Detail & Related papers  (2024-11-11T16:45:18Z)
• ReTok: Replacing Tokenizer to Enhance Representation Efficiency in Large Language Model [9.1108256816605]
  We propose a method to improve model representation and processing efficiency by replacing the tokenizers of large language models (LLMs) Our method can maintain the performance of the model after replacing the tokenizer, while significantly improving the decoding speed for long texts.
  arXiv  Detail & Related papers  (2024-10-06T03:01:07Z)
• Transformers to SSMs: Distilling Quadratic Knowledge to Subquadratic Models [92.36510016591782]
  We present a method that is able to distill a pretrained Transformer architecture into alternative architectures such as state space models (SSMs) Our method, called MOHAWK, is able to distill a Mamba-2 variant based on the Phi-1.5 architecture using only 3B tokens and a hybrid version (Hybrid Phi-Mamba) using 5B tokens. Despite using less than 1% of the training data typically used to train models from scratch, Phi-Mamba boasts substantially stronger performance compared to all past open-source non-Transformer models.
  arXiv  Detail & Related papers  (2024-08-19T17:48:11Z)
• Are Self-Attentions Effective for Time Series Forecasting? [4.990206466948269]
  Time series forecasting is crucial for applications across multiple domains and various scenarios. Recent findings have indicated that simpler linear models might outperform complex Transformer-based approaches. We introduce a new architecture, Cross-Attention-only Time Series transformer (CATS) Our model achieves superior performance with the lowest mean squared error and uses fewer parameters compared to existing models.
  arXiv  Detail & Related papers  (2024-05-27T06:49:39Z)
• MatFormer: Nested Transformer for Elastic Inference [94.1789252941718]
  MatFormer is a nested Transformer architecture designed to offer elasticity in a variety of deployment constraints. We show that a 2.6B decoder-only MatFormer language model (MatLM) allows us to extract smaller models spanning from 1.5B to 2.6B. We also observe that smaller encoders extracted from a universal MatFormer-based ViT (MatViT) encoder preserve the metric-space structure for adaptive large-scale retrieval.
  arXiv  Detail & Related papers  (2023-10-11T17:57:14Z)
• Understanding Parameter Sharing in Transformers [53.75988363281843]
  Previous work on Transformers has focused on sharing parameters in different layers, which can improve the performance of models with limited parameters by increasing model depth. We show that the success of this approach can be largely attributed to better convergence, with only a small part due to the increased model complexity. Experiments on 8 machine translation tasks show that our model achieves competitive performance with only half the model complexity of parameter sharing models.
  arXiv  Detail & Related papers  (2023-06-15T10:48:59Z)
• Scaling Pre-trained Language Models to Deeper via Parameter-efficient Architecture [68.13678918660872]
  We design a more capable parameter-sharing architecture based on matrix product operator (MPO) MPO decomposition can reorganize and factorize the information of a parameter matrix into two parts. Our architecture shares the central tensor across all layers for reducing the model size.
  arXiv  Detail & Related papers  (2023-03-27T02:34:09Z)
• Revision Transformers: Instructing Language Models to Change their Values [21.645935518842744]
  Current transformer language models (LM) are large-scale models with billions of parameters. We propose the Revision Transformer (RiT) to facilitate easy model updating. The specific combination of a large-scale pre-trained LM that inherently but also diffusely encodes world knowledge with a clear-structured revision engine makes it possible to update the model's knowledge with little effort and the help of user interaction.
  arXiv  Detail & Related papers  (2022-10-19T07:05:06Z)
• ClusTR: Exploring Efficient Self-attention via Clustering for Vision Transformers [70.76313507550684]
  We propose a content-based sparse attention method, as an alternative to dense self-attention. Specifically, we cluster and then aggregate key and value tokens, as a content-based method of reducing the total token count. The resulting clustered-token sequence retains the semantic diversity of the original signal, but can be processed at a lower computational cost.
  arXiv  Detail & Related papers  (2022-08-28T04:18:27Z)
• Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity [35.84448624327473]
  We simplify the MoE routing algorithm and design intuitive improved models with reduced communication and computational costs. We show large sparse models may be trained, for the first time, with lower precision (bfloat16) formats. We design models based off T5-Base and T5-Large to obtain up to 7x increases in pre-training speed with the same computational resources.
  arXiv  Detail & Related papers  (2021-01-11T16:11:52Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.