SOLAR: A State-Driven Online Learning Rate Scheduler for LLM Pretraining
cs.LG
Submitted: 2026-09-28
Updated: 2026-09-28
Sources
- DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model
- The Pile: An 800GB Dataset of Diverse Text for Language Modeling
- Using Deep Q-Learning to Control Optimization Hyperparameters
- Muon is Scalable for LLM Training
- An Empirical Model of Large-Batch Training
- VeLO: Training Versatile Learned Optimizers by Scaling Up
- Proximal Policy Optimization Algorithms
- Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism
- Spike No More: Stabilizing the Pre-training of Large Language Models
- Olmo 3
- Generalisable Agents for Neural Network Optimisation
- LLaMA: Open and Efficient Foundation Language Models
- Llama 2: Open Foundation and Fine-Tuned Chat Models
- Reinforcement Learning for Learning Rate Control
- Learning an Adaptive Learning Rate Schedule
- Qwen2 Technical Report
- OPT: Open Pre-trained Transformer Language Models
- Accelerating LLM Pre-Training through Flat-Direction Dynamics Enhancement
- ST-MoE: Designing Stable and Transferable Sparse Expert Models
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks