starAndHonor/flashmoon/examples/attn_cpu does not have a README file
Source Files
Install
Download zipPure-MoonBit WebGPU runtime and inference kernel library: WGSL kernels, 4D batched FlashAttention (MHA/GQA/MQA, causal, cross-attention), same API across wasm/native/js and WebGPU, with verified numerics.
webgpuattentionflashattentiongpuinference
Apache-2.0
Updated last month
34 downloads