Backward-State Policy Is Part of the Learning Algorithm
cs.LG
Submitted: 2026-09-30
Updated: 2026-09-30
Code: https://github.com/pytorch/pytorch
Sources
- GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints
- Inverse Path Tracing for Joint Material and Lighting Estimation
- The Mechanics of n-Player Differentiable Games
- Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation
- Unit Scaling: Out-of-the-Box Low-Precision Training
- Quartet: Native FP4 Training Can Be Optimal for Large Language Models
- FP4 All the Way: Fully Quantized Training of LLMs
- FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning
- FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness
- Scaling FP8 training to trillion-token LLMs
- Is Flash Attention Stable?
- The Llama 3 Herd of Models
- Towards Fully FP8 GEMM LLM Training at Scale
- Hardware-Aware FP4 FlashAttention-4
- MiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategies
- Scaling Laws for Precision
- Understanding Quantization-Aware Training: Gradients at Quantized Weights Bias to the Low-Loss Basin
- PRAC: Principal-Random Subspace for LLM Activation Compression and Memory-Efficient Training
- FP8 Formats for Deep Learning
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