OpenMachine-ai / transformer-tricksLinks
A collection of tricks and tools to speed up transformer models
☆169Updated 2 months ago
Alternatives and similar repositories for transformer-tricks
Users that are interested in transformer-tricks are comparing it to the libraries listed below
Sorting:
- ☆61Updated 5 months ago
- ☆237Updated 2 months ago
- Work in progress.☆72Updated last month
- Flash-Muon: An Efficient Implementation of Muon Optimizer☆160Updated 2 months ago
- Official repository for the paper "NeuZip: Memory-Efficient Training and Inference with Dynamic Compression of Neural Networks". This rep…☆59Updated 9 months ago
- Block Transformer: Global-to-Local Language Modeling for Fast Inference (NeurIPS 2024)☆160Updated 4 months ago
- Fused Qwen3 MoE layer for faster training, compatible with HF Transformers, LoRA, 4-bit quant, Unsloth☆167Updated this week
- RWKV-7: Surpassing GPT☆94Updated 9 months ago
- Chain of Experts (CoE) enables communication between experts within Mixture-of-Experts (MoE) models☆220Updated 2 months ago
- [ICLR 2025] Official PyTorch Implementation of Gated Delta Networks: Improving Mamba2 with Delta Rule☆199Updated 5 months ago
- Efficient LLM Inference over Long Sequences☆389Updated 2 months ago
- Towards Economical Inference: Enabling DeepSeek's Multi-Head Latent Attention in Any Transformer-based LLMs☆188Updated 2 months ago
- ☆80Updated 6 months ago
- Efficient triton implementation of Native Sparse Attention.☆208Updated 3 months ago
- Parallel Scaling Law for Language Model — Beyond Parameter and Inference Time Scaling☆432Updated 3 months ago
- ☆55Updated 3 months ago
- ☆144Updated 9 months ago
- Tree Attention: Topology-aware Decoding for Long-Context Attention on GPU clusters☆129Updated 8 months ago
- ☆123Updated 2 months ago
- Lightning Attention-2: A Free Lunch for Handling Unlimited Sequence Lengths in Large Language Models☆324Updated 6 months ago
- An efficient implementation of the NSA (Native Sparse Attention) kernel☆113Updated 2 months ago
- The evaluation framework for training-free sparse attention in LLMs☆90Updated 2 months ago
- Normalized Transformer (nGPT)☆186Updated 9 months ago
- Training-free Post-training Efficient Sub-quadratic Complexity Attention. Implemented with OpenAI Triton.☆145Updated this week
- RWKV is an RNN with transformer-level LLM performance. It can be directly trained like a GPT (parallelizable). So it's combining the best…☆51Updated 5 months ago
- ☆55Updated last year
- Repo for "LoLCATs: On Low-Rank Linearizing of Large Language Models"☆245Updated 6 months ago
- PyTorch implementation of models from the Zamba2 series.☆184Updated 7 months ago
- Official implementation of the paper: "ZClip: Adaptive Spike Mitigation for LLM Pre-Training".☆131Updated 2 weeks ago
- This repository contains code for the MicroAdam paper.☆19Updated 8 months ago