Mynah
Popular repositories Loading
-
mynah-slm
mynah-slm PublicFast native C inference engine for small language models, llama.cpp-style: streaming generation, thinking on/off, GGUF + safetensors, hand-written SIMD kernels, HTTP server with SSE. CPU-first and …
C
-
-
Repositories
- ingot Public
Zero-dependency C11 library for reading and writing GGUF and safetensors: zero-copy mmap, bit-exact decode of all 33 ggml quant types, runtime-dispatched SIMD kernels. Static lib or two-file amalgam.
- mynah-slm Public
Fast native C inference engine for small language models, llama.cpp-style: streaming generation, thinking on/off, GGUF + safetensors, hand-written SIMD kernels, HTTP server with SSE. CPU-first and no Python at runtime. Runs Qwen3 today, Gemma 4 next.
- mynah-app Public
- mynah-tts Public
Pure-C text-to-speech engine: autoregressive decoder + discrete neural audio codec, no Python at runtime. Sub-realtime on a 2020 M1 CPU (RTF 0.36 int8) and 0.26 on a mainstream NVIDIA GPU. f16/int8/int4 decode weights, memory-mapped, Metal & CUDA optional. Runs NVIDIA Magpie + NanoCodec, 12 languages.
- mynah-asr Public
Fast native C inference engine for speech recognition & translation, llama.cpp-style: streaming + offline ASR, word timestamps, int8/int4 quantization, CPU/Metal/CUDA. Today it runs the best open models (Parakeet, Canary, Nemotron) — built to host more engines tomorrow.
Top languages
Loading…
Most used topics
Loading…