Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
October 16, 2025ASEAN Journal of Scientific and Technological Reports

Scalable Deep Neural Network Training: Overcoming Memory Constraints with Performance Preservation

View Full Paper
Ask AI
Bookmark
Share

Authors

KRKaligoti RavikumarSri Venkateswara UniversityCSC. SivakumarSri Venkateswara University

Discussion

Loading...

Member takes

Implication

This method utilizes small-batch processing to achieve large-batch training, indicating improved resource efficiency.

Key Points

  • Training with large batch sizes is possible without sacrificing model performance, enhancing resource efficiency.
  • Small-batch processing uses a streaming mechanism and gradient accumulation to overcome memory constraints effectively.
  • Experiments show that this new method matches traditional approaches, confirming its viability for deep learning tasks.
  • The technique is especially beneficial for enhancing scalability in resource-limited environments.

Cite This Study

Ravikumar et al. (2025) studied this question.

synapsesocial.com/papers/68f0ba59c50c73ebef9faa5dhttps://doi.org/10.55164/ajstr.v28i5.256270
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Enabling Large Dynamic Neural Network Training with Learning-based Memory Management2024 · 11 citations
  2. 2Optimization Techniques For Large-Scale Deep Neural Networks: A Performance And Efficiency Analysis2024
  3. 3Mini-batch Coresets for Memory-efficient Language Model Training on Data Mixtures2024 · 1 citations
  4. 4DYNA-B: An Enhanced And Dynamic Batch Size Tuning For LSTM Neural Network2024
  5. 5Exploring Scalability of Deep Neural Networks for Real-Time Video Processing in Edge Devices2024