bstnxbt

dflash-mlx. Lossless DFlash speculative decoding for MLX on Apple Silicon

mola. Multi-LoRA inference server for Apple Silicon -- one base model, many adapters, zero reload

siliconrt. Experimental Apple Silicon inference substrate in C++ and Metal

solana-lora-bench. Fine-tuned LLM benchmark for Solana transaction generation. Dataset + LoRA + evaluation on Solana Bench.