juvi21 / CoPE-cudaLinks
Contextual Position Encoding but with some custom CUDA Kernels https://arxiv.org/abs/2405.18719
☆22Updated last year
Alternatives and similar repositories for CoPE-cuda
Users that are interested in CoPE-cuda are comparing it to the libraries listed below
Sorting:
- ☆55Updated 11 months ago
- An Experiment on Dynamic NTK Scaling RoPE☆64Updated last year
- Code for preprint "Metadata Conditioning Accelerates Language Model Pre-training (MeCo)"☆39Updated last month
- Implementation of the paper: "Leave No Context Behind: Efficient Infinite Context Transformers with Infini-attention" from Google in pyTO…☆55Updated 3 weeks ago
- Linear Attention Sequence Parallelism (LASP)☆84Updated last year
- triton ver of gqa flash attn, based on the tutorial☆11Updated 10 months ago
- ☆20Updated last year
- [ICML'24] The official implementation of “Rethinking Optimization and Architecture for Tiny Language Models”☆121Updated 5 months ago
- 32 times longer context window than vanilla Transformers and up to 4 times longer than memory efficient Transformers.☆48Updated 2 years ago
- Using FlexAttention to compute attention with different masking patterns☆44Updated 9 months ago
- DPO, but faster 🚀☆43Updated 6 months ago
- Here we will test various linear attention designs.☆59Updated last year
- ☆31Updated last year
- ☆48Updated last year
- The official implementation for Gated Attention for Large Language Models: Non-linearity, Sparsity, and Attention-Sink-Free☆44Updated last month
- ☆64Updated last year
- My Implementation of Q-Sparse: All Large Language Models can be Fully Sparsely-Activated☆32Updated 10 months ago
- Implementation of "LM-Infinite: Simple On-the-Fly Length Generalization for Large Language Models"☆41Updated 7 months ago
- Triton Implementation of HyperAttention Algorithm☆48Updated last year
- Official repository for the paper "Approximating Two-Layer Feedforward Networks for Efficient Transformers"☆38Updated 2 weeks ago
- Repository for Sparse Finetuning of LLMs via modified version of the MosaicML llmfoundry☆42Updated last year
- [NeurIPS 2023] Sparse Modular Activation for Efficient Sequence Modeling☆37Updated last year
- Reference implementation of "Softmax Attention with Constant Cost per Token" (Heinsen, 2024)☆24Updated last year
- This is a personal reimplementation of Google's Infini-transformer, utilizing a small 2b model. The project includes both model and train…☆57Updated last year
- Odysseus: Playground of LLM Sequence Parallelism☆70Updated last year
- GoldFinch and other hybrid transformer components☆45Updated 11 months ago
- Transformers at any scale☆41Updated last year
- Official repository of paper "RNNs Are Not Transformers (Yet): The Key Bottleneck on In-context Retrieval"☆27Updated last year
- Implementation of the model: "Reka Core, Flash, and Edge: A Series of Powerful Multimodal Language Models" in PyTorch☆30Updated 3 weeks ago
- ☆15Updated last week