DeepSeek-V3: Achieving Efficient LLM Scaling with 2,048 GPUs (arxiv.org) 7 pts| 10 months ago | 1 comment
