Lossless LLM compression for efficient GPU inference via dynamic-length float
Article URL: https://arxiv.org/abs/2504.11651 Comments URL: https://news.ycombinator.com/item?id=43796935 Points: 191 # Comments: 59

Article URL: https://arxiv.org/abs/2504.11651
Comments URL: https://news.ycombinator.com/item?id=43796935
Points: 191
# Comments: 59