Impact of the DeepSeek moment on inference compute скачать в хорошем качестве

Impact of the DeepSeek moment on inference compute

In this short talk, d-Matrix CTO Sudeep Bhoja discusses the release of the Deep Seek R1 model, highlighting its impact on inference compute. He discusses the evolution of reasoning models and the significance of inference time compute in enhancing model performance. He highlights the following: Reasoning models rely on “inference time compute”. They will unlock the golden age of inference DeepSeek R1 is only the first of many open models that will compete with frontier models. Distillation makes smaller models much more capable Unlocking efficiency from model architecture and algorithmic techniques today. Models are highly memory bound, so GPUs end up being under-utilized. Deploying with efficient inference compute platform like d-Matrix Corsair will result in faster speed, cost savings and energy efficiency. Sudeep explains the need for increased memory bandwidth to handle the generation of more tokens in reasoning models like Deep Seek R1, why GPUs will be inefficient and how d-Matrix Corsair with it’s unique memory-compute integration is ideal for such models.