NVIDIA / Megatron-Energon
Megatron's multi-modal data loader
β181Updated this week
Alternatives and similar repositories for Megatron-Energon:
Users that are interested in Megatron-Energon are comparing it to the libraries listed below
- π Efficiently (pre)training foundation models with native PyTorch features, including FSDP for training and SDPA implementation of Flashβ¦β232Updated 2 weeks ago
- ring-attention experimentsβ128Updated 5 months ago
- This repository contains the experimental PyTorch native float8 training UXβ222Updated 7 months ago
- Official repository for LightSeq: Sequence Level Parallelism for Distributed Training of Long Context Transformersβ207Updated 7 months ago
- Applied AI experiments and examples for PyTorchβ249Updated this week
- [ICLR 2025] COAT: Compressing Optimizer States and Activation for Memory-Efficient FP8 Trainingβ164Updated last month
- β158Updated last month
- [ICLR2025] Breaking Throughput-Latency Trade-off for Long Sequences with Speculative Decodingβ110Updated 3 months ago
- π Collection of components for development, training, tuning, and inference of foundation models leveraging PyTorch native components.β189Updated this week
- USP: Unified (a.k.a. Hybrid, 2D) Sequence Parallel Attention for Long Context Transformers Model Training and Inferenceβ452Updated this week
- Tritonbench is a collection of PyTorch custom operators with example inputs to measure their performance.β104Updated this week
- Efficient LLM Inference over Long Sequencesβ365Updated last month
- π₯ A minimal training framework for scaling FLA modelsβ82Updated this week
- Odysseus: Playground of LLM Sequence Parallelismβ66Updated 9 months ago
- Load compute kernels from the Hubβ99Updated this week
- LLM KV cache compression made easyβ440Updated this week
- Block Transformer: Global-to-Local Language Modeling for Fast Inference (NeurIPS 2024)β150Updated 3 months ago
- Implementation of π Ring Attention, from Liu et al. at Berkeley AI, in Pytorchβ506Updated 5 months ago
- Zero Bubble Pipeline Parallelismβ373Updated 3 weeks ago
- β74Updated 3 months ago
- Triton-based implementation of Sparse Mixture of Experts.β208Updated 3 months ago
- [ICLR 2025] DuoAttention: Efficient Long-Context LLM Inference with Retrieval and Streaming Headsβ438Updated last month
- PyTorch bindings for CUTLASS grouped GEMM.β74Updated 4 months ago
- PyTorch bindings for CUTLASS grouped GEMM.β108Updated 2 months ago
- Fast low-bit matmul kernels in Tritonβ267Updated this week
- Scalable and Performant Data Loadingβ230Updated this week
- β116Updated last year
- PyTorch per step fault tolerance (actively under development)β267Updated this week
- Explorations into some recent techniques surrounding speculative decodingβ250Updated 3 months ago
- [ICML 2024] Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inferenceβ259Updated 4 months ago