Threshold Filtering Packing for Supervised Fine-Tuning: Training Related Samples within Packs
Packing for Supervised Fine-Tuning (SFT) in autoregressive models involves concatenating data points of varying lengths until reaching the designed maximum length to facilitate GPU processing. However, randomly concatenating data points and feeding them into an autoregressive transformer can lead to cross-contamination of sequences due to the significant difference in their subject matter. The mainstream approaches in SFT ensure that each token in the attention calculation phase only focuses on tokens within its own short sequence, without providing additional learning signals for the preceding context. To address these challenges, we introduce Threshold Filtering Packing (TFP), a method that selects samples with related context while maintaining sufficient diversity within the same pack. Our experiments show that TFP offers a simple-to-implement and scalable approach that significantly enhances SFT performance, with observed improvements of up to 7\% on GSM8K, 4\% on HumanEval, and 15\% on the adult-census-income dataset.
Code (0)
등록된 구현이 없습니다.
Tasks
DiversityGPUGSM8KHumanEvalMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Packing Analysis: Packing Is More Appropriate for Large Models or Datasets in Supervised Fine-tuning
Packing, initially utilized in the pre-training phase, is an optimization technique designed to maximize hardware resource efficiency by combining different training sequences to fit the model's maximum input length. Alt…
Open-Qwen2VL: Compute-Efficient Pre-Training of Fully-Open Multimodal LLMs on Academic Resources
The reproduction of state-of-the-art multimodal LLM pre-training faces barriers at every stage of the pipeline, including high-quality data filtering, multimodal data mixture strategies, sequence packing techniques, and …
GPULarge Language ModelMultimodal Large Language ModelFeedback-Driven Pseudo-Label Reliability Assessment: Redefining Thresholding for Semi-Supervised Semantic Segmentation
Semi-supervised learning leverages unlabeled data to enhance model performance, addressing the limitations of fully supervised approaches. Among its strategies, pseudo-supervision has proven highly effective, typically r…
Pseudo LabelPseudo Label FilteringSemantic SegmentationSemi-Supervised Semantic SegmentationNear-Policy: Accelerating On-Policy Distillation via Asynchronous Generation and Selective Packing
Standard knowledge distillation for autoregressive models often suffers from distribution mismatch. While on-policy methods mitigate this by leveraging student-generated outputs, they rely on computationally expensive Re…
Knowledge DistillationReinforcement LearningPreference Packing: Efficient Preference Optimization for Large Language Models
Resource-efficient training optimization techniques are becoming increasingly important as the size of large language models (LLMs) continues to grow. In particular, batch packing is commonly used in pre-training and sup…