Speculative Decoding Efficiency
Methods for improving the speed and memory efficiency of speculative decoding in language models, especially at long sequence lengths.
1 papers
Where this stands
The written synthesis of this thread is for subscribers. Subscribe.