Tuning Language Models by Mixture-of-Depths Ensemble

• URL: http://arxiv.org/abs/2410.13077v1
• Date: Wed, 16 Oct 2024 22:51:45 GMT
• Title: Tuning Language Models by Mixture-of-Depths Ensemble
• Authors: Haoyan Luo, Lucia Specia,
• Abstract summary: Transformer-based Large Language Models (LLMs) traditionally rely on final-layer loss for training and final-layer representations for predictions. We find that focusing training efforts on intermediate layers can yield training losses comparable to those of final layers. We introduce a novel tuning framework, Mixture-of-Depths (MoD), which trains late layers as ensembles contributing to the final logits.
• Score: 23.10522891268232
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Transformer-based Large Language Models (LLMs) traditionally rely on final-layer loss for training and final-layer representations for predictions, potentially overlooking the predictive power embedded in intermediate layers. Surprisingly, we find that focusing training efforts on these intermediate layers can yield training losses comparable to those of final layers, with complementary test-time performance. We introduce a novel tuning framework, Mixture-of-Depths (MoD), which trains late layers as ensembles contributing to the final logits through learned routing weights. With the auxiliary distillation loss and additional normalization modules, we ensure that the outputs of the late layers adapt to language modeling. Our MoD framework, which can be integrated with any existing tuning method, shows consistent improvement on various language modelling tasks. Furthermore, by replacing traditional trainable modules with MoD, our approach achieves similar performance with significantly fewer trainable parameters, demonstrating the potential of leveraging predictive power from intermediate representations during training.

Related papers

• DiRL: An Efficient Post-Training Framework for Diffusion Language Models [54.405206032785706]
  Diffusion Language Models (dLLMs) have emerged as promising alternatives to Auto-Regressive (AR) models.<n>Existing methods suffer from computational inefficiency and objective mismatches between training and inference.<n>We introduce DiRL, an efficient post-training framework that tightly integrates FlexAttention-accelerated blockwise training with LMDeploy-optimized inference.
  arXiv  Detail & Related papers  (2025-12-23T08:33:19Z)
• Similarity-Distance-Magnitude Language Models [0.0]
  We introduce Similarity-Distance-Magnitude (SDM) language models (LMs)<n>LMs are sequence prediction models fine-tuned to maximize the proportion of generations in the well-calibrated, high-probability region partitioned by a final-layer SDM activation layer used for binary classification of instruction-following.
  arXiv  Detail & Related papers  (2025-10-30T06:42:15Z)
• Platonic Grounding for Efficient Multimodal Language Models [22.715168904364756]
  We motivate and propose a simple modification to existing multimodal frameworks that rely on aligning pretrained models. Our work also has implications for combining pretrained models into larger systems efficiently.
  arXiv  Detail & Related papers  (2025-04-27T18:56:26Z)
• Meta-Learning Adaptable Foundation Models [37.458141335750696]
  We introduce a meta-learning framework infused with PEFT in this intermediate retraining stage to learn a model that can be easily adapted to unseen tasks. In this setting, we demonstrate the suboptimality of standard retraining for finding an adaptable set of parameters. We then apply these theoretical insights to retraining the RoBERTa model to predict the continuation of conversations within the ConvAI2 dataset.
  arXiv  Detail & Related papers  (2024-10-29T17:24:18Z)
• Transferable Post-training via Inverse Value Learning [83.75002867411263]
  We propose modeling changes at the logits level during post-training using a separate neural network (i.e., the value network) After training this network on a small base model using demonstrations, this network can be seamlessly integrated with other pre-trained models during inference. We demonstrate that the resulting value network has broad transferability across pre-trained models of different parameter sizes.
  arXiv  Detail & Related papers  (2024-10-28T13:48:43Z)
• Joint Fine-tuning and Conversion of Pretrained Speech and Language Models towards Linear Complexity [11.302828987873497]
  We present a Cross-Architecture Layerwise Distillation (CALD) approach that jointly converts a transformer model to a linear time substitute and fine-tunes it to a target task. We show that CALD can effectively recover the result of the original model, and that the guiding strategy contributes to the result.
  arXiv  Detail & Related papers  (2024-10-09T13:06:43Z)
• Boosting Continual Learning of Vision-Language Models via Mixture-of-Experts Adapters [65.15700861265432]
  We present a parameter-efficient continual learning framework to alleviate long-term forgetting in incremental learning with vision-language models. Our approach involves the dynamic expansion of a pre-trained CLIP model, through the integration of Mixture-of-Experts (MoE) adapters. To preserve the zero-shot recognition capability of vision-language models, we introduce a Distribution Discriminative Auto-Selector.
  arXiv  Detail & Related papers  (2024-03-18T08:00:23Z)
• Expedited Training of Visual Conditioned Language Generation via Redundancy Reduction [61.16125290912494]
  $textEVL_textGen$ is a framework designed for the pre-training of visually conditioned language generation models. We show that our approach accelerates the training of vision-language models by a factor of 5 without a noticeable impact on overall performance.
  arXiv  Detail & Related papers  (2023-10-05T03:40:06Z)
• Domain Generalization via Balancing Training Difficulty and Model Capability [61.053202176230904]
  Domain generalization (DG) aims to learn domain-generalizable models from one or multiple source domains that can perform well in unseen target domains. Despite its recent progress, most existing work suffers from the misalignment between the difficulty level of training samples and the capability of contemporarily trained models. We design MoDify, a Momentum Difficulty framework that tackles the misalignment by balancing the seesaw between the model's capability and the samples' difficulties.
  arXiv  Detail & Related papers  (2023-09-02T07:09:23Z)
• Adapted Multimodal BERT with Layer-wise Fusion for Sentiment Analysis [84.12658971655253]
  We propose Adapted Multimodal BERT, a BERT-based architecture for multimodal tasks. adapter adjusts the pretrained language model for the task at hand, while the fusion layers perform task-specific, layer-wise fusion of audio-visual information with textual BERT representations. In our ablations we see that this approach leads to efficient models, that can outperform their fine-tuned counterparts and are robust to input noise.
  arXiv  Detail & Related papers  (2022-12-01T17:31:42Z)
• End-to-End Training for Back-Translation with Categorical Reparameterization Trick [0.0]
  Back-translation is an effective semi-supervised learning framework in neural machine translation (NMT) A pre-trained NMT model translates monolingual sentences and makes synthetic bilingual sentence pairs for the training of the other NMT model. The discrete property of translated sentences prevents information gradient from flowing between the two NMT models.
  arXiv  Detail & Related papers  (2022-02-17T06:31:03Z)
• Pretrained Transformers as Universal Computation Engines [105.00539596788127]
  We investigate the capability of a transformer pretrained on natural language to generalize to other modalities with minimal finetuning. We study finetuning it on a variety of sequence classification tasks spanning numerical computation, vision, and protein fold prediction. We find that such pretraining enables FPT to generalize in zero-shot to these modalities, matching the performance of a transformer fully trained on these tasks.
  arXiv  Detail & Related papers  (2021-03-09T06:39:56Z)
• Rethinking embedding coupling in pre-trained language models [46.11201932668366]
  We re-evaluate the standard practice of sharing weights between input and output embeddings in pre-trained language models. We show that decoupled embeddings provide increased modeling flexibility, allowing us to significantly improve the efficiency of parameter allocation. We are able to train models that achieve strong performance on the XTREME benchmark without increasing the number of parameters at the fine-tuning stage.
  arXiv  Detail & Related papers  (2020-10-24T07:43:00Z)

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.