arXiv Machine Learning
Jul 3
DecompRL: Solving Harder Problems by Learning Modular Code Generation
arXiv:2607. 02390v1 Announce Type: new Abstract: How can Large Language Models (LLMs) solve problems they currently cannot?
How to Train Your Little Dragon: Teaching a Small Language Model to Construct and Refine GPU Optimization Scripts
The paper introduces HOW TO TRAIN YOUR LITTLE DRAGON, a method for teaching a small language model to generate and refine GPU optimization scripts. Using a compact, lossless representation of compiler traces, the authors train the model on 108,594 examples derived from TVM MetaSchedule, first with supervised fine‑tuning and then with reinforcement learning that rewards correctness and performance. After training, an 8‑billion‑parameter model achieves over 91% correctness in both construction and refinement tasks and outperforms several much larger LLMs on kernel performance.
Related stories
arXiv Machine Learning
Jul 10
Selective Left-Shift: Turning Test-Time Compute and Difficulty-based Curation into Training Data for Low-Resource Code Generation
arXiv:2607. 07748v1 Announce Type: new Abstract: Large Language Models achieve strong code generation for high resource languages like Python and Java but suffer sharp performance drops on Low-Resource Programming Languages~(LRPLs) such as Julia.
arXiv Machine Learning
Jul 7
NKI-Agent: Domain-Specific Fine-Tuning and Agentic Tool Use for Neuron Kernel Generation
arXiv:2607. 04395v1 Announce Type: new Abstract: Recent agentic approaches to LLM-based kernel generation have achieved impressive results on CUDA.
arXiv AI
Jul 24
Multi-turn RL with Structural and Performance Aware Rewards for CUDA Kernel Generation
arXiv:2607. 20908v1 Announce Type: cross Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a powerful technique to enhance the reasoning capacity of LLMs for optimized code generation.
arXiv AI
Aug 5
On the Limits of Layer Pruning for Generative Reasoning in Large Language Models
arXiv:2602. 01997v3 Announce Type: replace-cross Abstract: Recent work has shown that layer pruning can effectively compress large language models (LLMs) while retaining strong performance on classification benchmarks, often with little or no finetuning.
arXiv Machine Learning
Aug 4
LEAP: Lean Environment-Feedback via Adaptive Pruning for Code RL in GPU Kernel Generation
arXiv:2608. 01804v1 Announce Type: new Abstract: Post-training large language models (LLMs) via reinforcement learning (RL) has significantly advanced code generation capabilities.
Hugging Face Trending Papers
Jul 23
Multi-turn RL with Structural and Performance Aware Rewards for CUDA Kernel Generation
Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a powerful technique to enhance the reasoning capacity of LLMs for optimized code generation. However, existing RLVR approaches primarily rely on outcome-based signals such as correctness and speedup, overlooking performance-critical structural properties of programs that are essential for generating optimized code.
arXiv Machine Learning
Jun 4
MusaCoder: Native GPU Kernel Generation with Full-Stack Training on Moore Threads GPU
arXiv:2606. 04847v1 Announce Type: cross Abstract: Native GPU kernel generation turns high-level tensor programs into executable, efficient low-level code.
arXiv AI
23h ago
Prompt Codebooks: Discrete Compositional Optimization for Language Model Instruction Refinement
arXiv:2605.28360v3 Announce Type: replace Abstract: Automatic prompt optimization (APO) has driven significant gains in LLM-based agentic workflows. However, most existing methods treat each task's p...
arXiv Machine Learning
Sep 1
Agnostics: Learning to Code in Any Programming Language via Reinforcement with a Universal Learning Environment
Agnostics is a language‑agnostic post‑training pipeline that uses reinforcement learning with verifiable rewards (RLVR) to improve large language models on low‑resource programming languages. By rewriting unit‑test datasets into a language‑independent I/O format, providing a short configuration for compiling and running code, and employing a single verifier that judges code by observable behavior, Agnostics eliminates the need for language‑specific engineering. Applied to Lua, Julia, R, OCaml, and Fortran, it boosts Qwen‑3 4B to rival larger models, scales to diverse families, and achieves new state‑of‑the‑art pass@1 on MultiPL‑E and a new multi‑language LiveCodeBench.
arXiv AI
Sep 16
Ave: Guiding Agentic GPU Optimization Using Data-Flow Invariants
arXiv:2604.18616v2 Announce Type: replace-cross Abstract: LLM coding agents can generate correct GPU kernels, but their performance still trails expert libraries. Reaching peak throughput requires co...
arXiv Machine Learning
Jun 3
KForge: LLM-Driven Cross-Platform Kernel Generation for AI Accelerators
arXiv:2606. 02963v1 Announce Type: new Abstract: Production inference increasingly targets a heterogeneous mix of accelerators.