Summarization on CNN/Daily Mail 19 (test)
1.45Speedup (x)SDFP
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SDFPModel=LLaMA-2-70B, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 1.45 | 0.96 | |
| SWIFTModel=LLaMA-2-70B, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 1.43 | 0.99 | |
| SWIFTModel=LLaMA-2-13B, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 1.37 | 0.99 | |
| SDFPModel=LLaMA-2-13B, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 1.32 | 0.95 | |
| SWIFTModel=LLaMA-2-13B-Chat, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 1.28 | 0.9 | |
| SDFPModel=LLaMA-2-13B-Chat, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 1.23 | 0.92 | |
| Lookahead DecodingModel=LLaMA-2-13B, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 1.16 | — | |
| Lookahead DecodingModel=LLaMA-2-13B-Chat, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 1.15 | — | |
| Lookahead DecodingModel=LLaMA-2-70B, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 1.15 | — | |
| VANILLAModel=LLaMA-2-13B, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 1 | — | |
| VANILLAModel=LLaMA-2-13B-Chat, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 1 | — | |
| VANILLAModel=LLaMA-2-70B, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 1 | — | |
| Parallel DecodingModel=LLaMA-2-13B-Chat, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 0.96 | — | |
| Parallel DecodingModel=LLaMA-2-13B, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 0.95 | — | |
| Parallel DecodingModel=LLaMA-2-70B, Decoding Strategy=Greedy, Batch size=1, Shot configuration=1-shot, Maximum output length=642026.02 | 0.95 | — |