ROCm / xformersLinks
Hackable and optimized Transformers building blocks, supporting a composable construction.
☆34Updated 2 weeks ago
Alternatives and similar repositories for xformers
Users that are interested in xformers are comparing it to the libraries listed below
Sorting:
- Fast and memory-efficient exact attention☆207Updated this week
- AMD related optimizations for transformer models☆96Updated 2 months ago
- [WIP] Better (FP8) attention for Hopper☆32Updated 10 months ago
- 8-bit CUDA functions for PyTorch☆69Updated 3 months ago
- A high-throughput and memory-efficient inference and serving engine for LLMs☆113Updated this week
- ☆55Updated this week
- Development repository for the Triton language and compiler☆140Updated this week
- ☆79Updated last year
- A Suite for Parallel Inference of Diffusion Transformers (DiTs) on multi-GPU Clusters☆53Updated last year
- Ahead of Time (AOT) Triton Math Library☆85Updated 3 weeks ago
- ☆78Updated last year
- [DEPRECATED] Moved to ROCm/rocm-libraries repo☆113Updated this week
- ☆71Updated 9 months ago
- a simple Flash Attention v2 implementation with ROCM (RDNA3 GPU, roc wmma), mainly used for stable diffusion(ComfyUI) in Windows ZLUDA en…☆50Updated last year
- (WIP) Parallel inference for black-forest-labs' FLUX model.☆18Updated last year
- A forked version of flux-fast that makes flux-fast even faster with cache-dit, 3.3x speedup on NVIDIA L20.☆24Updated 5 months ago
- Kernel Library Wheel for SGLang☆17Updated this week
- Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more☆24Updated this week
- PyTorch half precision gemm lib w/ fused optional bias + optional relu/gelu☆76Updated last year
- ☆162Updated 6 months ago
- ☆116Updated 7 months ago
- It is an LLM-based AI agent, which can write correct and efficient gpu kernels automatically.☆53Updated this week
- The AMD rocAL is designed to efficiently decode and process images and videos from a variety of storage formats and modify them through a…☆24Updated 3 weeks ago
- DeeperGEMM: crazy optimized version☆74Updated 8 months ago
- A safetensors extension to efficiently store sparse quantized tensors on disk☆228Updated this week
- Quantize transformers to any learned arbitrary 4-bit numeric format☆50Updated 6 months ago
- QuIP quantization☆61Updated last year
- Framework to reduce autotune overhead to zero for well known deployments.☆92Updated 3 months ago
- A Python library transfers PyTorch tensors between CPU and NVMe☆123Updated last year
- ☆22Updated 2 months ago