Related papers: Well-Read Students Learn Better: On the Importance…
Multilingual models are often particularly dependent on scaling to generalize to a growing number of languages. Compression techniques are widely relied upon to reconcile the growth in model size with real world resource constraints, but…
Deep learning based models are relatively large, and it is hard to deploy such models on resource-limited devices such as mobile phones and embedded devices. One possible solution is knowledge distillation whereby a smaller model (student…
Common knowledge distillation methods require the teacher model and the student model to be trained on the same task. However, the usage of embeddings as teachers has also been proposed for different source tasks and target tasks. Prior…
Language models are pre-trained using large corpora of generic data like book corpus, common crawl and Wikipedia, which is essential for the model to understand the linguistic characteristics of the language. New studies suggest using…
In self-supervised learning, self-distilled methods have shown impressive performance, learning representations useful for downstream tasks and even displaying emergent properties. However, state-of-the-art methods usually rely on ensembles…
Knowledge distillation transfers the knowledge from a cumbersome teacher to a small student. Recent results suggest that the student-friendly teacher is more appropriate to distill since it provides more transferable knowledge. In this…
As pretrained transformer language models continue to achieve state-of-the-art performance, the Natural Language Processing community has pushed for advances in model compression and efficient attention mechanisms to address high…
Pre-trained language models such as BERT have shown remarkable effectiveness in various natural language processing tasks. However, these models usually contain millions of parameters, which prevents them from practical deployment on…
This paper addresses the limitations of large-scale language models in safety alignment and robustness by proposing a fine-tuning method that combines contrastive distillation with noise-robust training. The method freezes the backbone…
Multilingual models have been widely used for cross-lingual transfer to low-resource languages. However, the performance on these languages is hindered by their underrepresentation in the pretraining data. To alleviate this problem, we…
Language modeling on large-scale datasets leads to impressive performance gains on various downstream language tasks. The validation pre-training loss (or perplexity in autoregressive language modeling) is often used as the evaluation…
Knowledge distillation compresses a larger neural model (teacher) into smaller, faster student models by training them to match teacher outputs. However, the internal computational transformations that occur during this process remain…
Existing methods for distillation do not efficiently utilize the training data. This work presents a novel approach to perform distillation using only a subset of the training data, making it more data-efficient. For this purpose, the…
The remarkable performance gains realized by large pretrained models, e.g., GPT-3, hinge on the massive amounts of data they are exposed to during training. Analogously, distilling such large models to compact models for efficient…
Knowledge distillation (KD) is commonly deemed as an effective model compression technique in which a compact model (student) is trained under the supervision of a larger pretrained model or an ensemble of models (teacher). Various…
Knowledge distillation generally assumes a strong-to-weak relationship where stronger teachers yield better students. In this work, we examine this assumption about distillation in large language model pretraining. By varying architecture…
Model distillation aims to distill the knowledge of a complex model into a simpler one. In this paper, we consider an alternative formulation called dataset distillation: we keep the model fixed and instead attempt to distill the knowledge…
Language models significantly benefit from context tokens, such as prompts or scratchpads. They perform better when prompted with informative instructions, and they acquire new reasoning capabilities by generating a scratch-pad before…
Knowledge distillation is the process of transferring the knowledge from a large model to a small model. In this process, the small model learns the generalization ability of the large model and retains the performance close to that of the…
The extra trust brought by the model interpretation has made it an indispensable part of machine learning systems. But to explain a distilled model's prediction, one may either work with the student model itself, or turn to its teacher…