G github.com Fmllm: 4mb training data, 100mb model, Fibonacci embeddings, near-coherent. WTF? 37 points gabriel666smith 1 year ago 27 comments
