latentCall145 / channels-last-groupnormLinks
A CUDA kernel for NHWC GroupNorm for PyTorch
☆19Updated 7 months ago
Alternatives and similar repositories for channels-last-groupnorm
Users that are interested in channels-last-groupnorm are comparing it to the libraries listed below
Sorting:
- Decoding Attention is specially optimized for MHA, MQA, GQA and MLA using CUDA core for the decoding stage of LLM inference.☆38Updated 2 weeks ago
- Multiple GEMM operators are constructed with cutlass to support LLM inference.☆18Updated 9 months ago
- Awesome code, projects, books, etc. related to CUDA☆17Updated last week
- ☆29Updated 4 months ago
- 使用 cutlass 仓库在 ada 架构上实现 fp8 的 flash attention☆71Updated 10 months ago
- Quantized Attention on GPU☆44Updated 7 months ago
- ☆60Updated 2 months ago
- GPTQ inference TVM kernel☆40Updated last year
- ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.☆80Updated last month
- Standalone Flash Attention v2 kernel without libtorch dependency☆110Updated 9 months ago
- An auxiliary project analysis of the characteristics of KV in DiT Attention.☆31Updated 6 months ago
- (WIP) Parallel inference for black-forest-labs' FLUX model.☆19Updated 7 months ago
- [WIP] Better (FP8) attention for Hopper☆30Updated 4 months ago
- ☆11Updated 3 months ago
- ☆49Updated last month
- A practical way of learning Swizzle☆20Updated 4 months ago
- ☆48Updated last month
- OneFlow Serving☆20Updated 2 months ago
- ☆14Updated 10 months ago
- A Suite for Parallel Inference of Diffusion Transformers (DiTs) on multi-GPU Clusters☆47Updated 11 months ago
- DeeperGEMM: crazy optimized version☆69Updated last month
- ☆19Updated 8 months ago
- ☆96Updated 9 months ago
- Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.☆38Updated 4 months ago
- Triton adapter for Ascend. Mirror of https://gitee.com/ascend/triton-ascend☆54Updated this week
- TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing.☆90Updated 2 weeks ago
- A standalone GEMM kernel for fp16 activation and quantized weight, extracted from FasterTransformer☆92Updated 3 weeks ago
- Implement Flash Attention using Cute.☆87Updated 6 months ago
- Odysseus: Playground of LLM Sequence Parallelism☆70Updated last year
- A GPU-optimized system for efficient long-context LLMs decoding with low-bit KV cache.☆47Updated 2 weeks ago