Remix.run Logo
xnx 2 hours ago

Is the diffusion approach any use in Multi-Token Prediction (MTP) drafters? https://blog.google/innovation-and-ai/technology/developers-...

fcanesin 2 hours ago | parent | next [-]

Yes, DFlash is currently a SOTA speculative decoding method that Xiaomi just used in their MiMo model for >1000tkps

doctorpangloss 2 hours ago | parent | prev [-]

MTP is a training optimization. Drafting requires verification, and verification is the full model inference. Speculative decoders are the name for the inference time optimization, that is more like a verifier that is a smaller model.