Portrait de Kshitij Gupta n'est pas disponible

Kshitij Gupta

kshitij.gupta@mila.quebec
Collaborateur·rice alumni - Université de Montréal
Superviseur⋅e principal⋅e
Irina Rish
Co-superviseur⋅e
Sarath Chandar Anbil Parthipan

Publications

Simple and Scalable Strategies to Continually Pre-train Large Language Models
Adam Ibrahim
Benjamin Thérien
Kshitij Gupta
Mats Leon Richter
Quentin Anthony
Timothee LESORT
Eugene Belilovsky
Irina Rish
Large language models (LLMs) are routinely pre-trained on billions of tokens, only to start the process over again once new data becomes ava… (voir plus)ilable. A much more efficient solution is to continually pre-train these models, saving significant compute compared to re-training. However, the distribution shift induced by new data typically results in degraded performance on previous data or poor adaptation to the new data. In this work, we show that a simple and scalable combination of learning rate (LR) re-warming, LR re-decaying, and replay of previous data is sufficient to match the performance of fully re-training from scratch on all available data, as measured by the final loss and the average score on several language model (LM) evaluation benchmarks. Specifically, we show this for a weak but realistic distribution shift between two commonly used LLM pre-training datasets (English
2024-03-13
ArXiv (prépublication)
doi.org
arxiv.org
Continual Pre-Training of Large Language Models: How to (re)warm your model?
Kshitij Gupta
Benjamin Thérien
Adam Ibrahim
Mats Leon Richter
Quentin Gregory Anthony
Eugene Belilovsky
Irina Rish
Timothee LESORT
Large language models (LLMs) are routinely pre-trained on billions of tokens, only to restart the process over again once new data becomes a… (voir plus)vailable. A much cheaper and more efficient solution would be to enable the continual pre-training of these models, i.e. updating pre-trained models with new data instead of re-training them from scratch. However, the distribution shift induced by novel data typically results in degraded performance on past data. Taking a step towards efficient continual pre-training, in this work, we examine the effect of different warm-up strategies. Our hypothesis is that the learning rate must be re-increased to improve compute efficiency when training on a new dataset. We study the warmup phase of models pre-trained on the Pile (upstream data, 300B tokens) as we continue to pre-train on SlimPajama (downstream data, 297B tokens), following a linear warmup and cosine decay schedule. We conduct all experiments on the Pythia 410M language model architecture and evaluate performance through validation perplexity. We experiment with different pre-training checkpoints, various maximum learning rates, and various warmup lengths. Our results show that while rewarming models first increases the loss on upstream and downstream data, in the longer run it improves the downstream performance, outperforming models trained from scratch
2023-06-20
ICML.cc/2023/Workshop/ES-FoMO (poster)
doi.org
openreview.net
Broken Neural Scaling Laws
Ethan Caballero
Kshitij Gupta
Irina Rish
David Scott Krueger
We present a smoothly broken power law functional form (that we refer to as a Broken Neural Scaling Law (BNSL)) that accurately models&extra… (voir plus)polates the scaling behaviors of deep neural networks (i.e. how the evaluation metric of interest varies as amount of compute used for training (or inference), number of model parameters, training dataset size, model input size, number of training steps, or upstream performance varies) for various architectures&for each of various tasks within a large&diverse set of upstream&downstream tasks, in zero-shot, prompted,&finetuned settings. This set includes large-scale vision, language, audio, video, diffusion, generative modeling, multimodal learning, contrastive learning, AI alignment, AI capabilities, robotics, out-of-distribution (OOD) generalization, continual learning, transfer learning, uncertainty estimation / calibration, OOD detection, adversarial robustness, distillation, sparsity, retrieval, quantization, pruning, fairness, molecules, computer programming/coding, math word problems,"emergent phase transitions", arithmetic, supervised learning, unsupervised/self-supervised learning,&reinforcement learning (single agent&multi-agent). When compared to other functional forms for neural scaling, this functional form yields extrapolations of scaling behavior that are considerably more accurate on this set. Moreover, this functional form accurately models&extrapolates scaling behavior that other functional forms are incapable of expressing such as the nonmonotonic transitions present in the scaling behavior of phenomena such as double descent&the delayed, sharp inflection points present in the scaling behavior of tasks such as arithmetic. Lastly, we use this functional form to glean insights about the limit of the predictability of scaling behavior. Code is available at https://github.com/ethancaballero/broken_neural_scaling_laws
2023-02-01
ICLR.cc/2023/Conference (poster)
doi.org
openreview.net