RLsys-Foundation / APRILLinks
APRIL: Active Partial Rollouts in Reinforcement Learning to Tame Long-tail Generation. A system-level optimization for scalable LLM training.
☆44Updated last month
Alternatives and similar repositories for APRIL
Users that are interested in APRIL are comparing it to the libraries listed below
Sorting:
- 🔥 LLM-powered GPU kernel synthesis: Train models to convert PyTorch ops into optimized Triton kernels via SFT+RL. Multi-turn compilation…☆99Updated 2 weeks ago
- Accelerate LLM preference tuning via prefix sharing with a single line of code☆51Updated 4 months ago
- ☆143Updated last week
- Odysseus: Playground of LLM Sequence Parallelism☆78Updated last year
- Implementation for FP8/INT8 Rollout for RL training without performence drop.☆275Updated 2 weeks ago
- [NeurIPS 2024] Efficient LLM Scheduling by Learning to Rank☆64Updated last year
- Official Implementation of SAM-Decoding: Speculative Decoding via Suffix Automaton☆36Updated 9 months ago
- LongSpec: Long-Context Lossless Speculative Decoding with Efficient Drafting and Verification☆67Updated 4 months ago
- ☆81Updated 7 months ago
- 16-fold memory access reduction with nearly no loss☆107Updated 7 months ago
- Efficient Reinforcement Learning for Language Models☆43Updated this week
- A lightweight reinforcement learning framework that integrates seamlessly into your codebase, empowering developers to focus on algorithm…☆81Updated 2 months ago
- PoC for "SpecReason: Fast and Accurate Inference-Time Compute via Speculative Reasoning" [NeurIPS '25]☆58Updated last month
- ☆46Updated 6 months ago
- Code for paper: [ICLR2025 Oral] FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference☆154Updated last month
- [ICLR 2025] TidalDecode: A Fast and Accurate LLM Decoding with Position Persistent Sparse Attention☆49Updated 3 months ago
- ☆50Updated 6 months ago
- [NeurIPS 2025] Scaling Speculative Decoding with Lookahead Reasoning☆52Updated 3 weeks ago
- [ICLR2025] Breaking Throughput-Latency Trade-off for Long Sequences with Speculative Decoding☆131Updated 11 months ago
- ☆132Updated 5 months ago
- An efficient implementation of the NSA (Native Sparse Attention) kernel☆126Updated 5 months ago
- The official implementation of paper: SimLayerKV: A Simple Framework for Layer-Level KV Cache Reduction.☆50Updated last year
- [ICLR 2025] PEARL: Parallel Speculative Decoding with Adaptive Draft Length☆130Updated 3 weeks ago
- GEAR: An Efficient KV Cache Compression Recipefor Near-Lossless Generative Inference of LLM☆170Updated last year
- ☆120Updated 5 months ago
- A simple calculation for LLM MFU.☆50Updated 2 months ago
- [ICML 2025] SparseLoRA: Accelerating LLM Fine-Tuning with Contextual Sparsity☆60Updated 4 months ago
- ☆34Updated 9 months ago
- ☆20Updated 8 months ago
- Pytorch implementation of our paper accepted by ICML 2024 -- CaM: Cache Merging for Memory-efficient LLMs Inference☆47Updated last year