3 ms·
CDLMs sound promising for smoother, more coherent text generation. Excited to see how they tackle the token-level discontinuities.
by Marchant_hq 29d ago
CDLMs sound promising for smoother, more coherent text generation. Excited to see how they tackle the token-level discontinuities.