TensorRT-LLMs/tests/integration/defs/accuracy/references/json_mode_eval.yaml
Enwei Zhu 5ff3a65b23
[TRTLLM-7028][feat] Enable guided decoding with speculative decoding (part 2: one-model engine) (#6948)
Signed-off-by: Enwei Zhu <21126786+syuoni@users.noreply.github.com>
2025-09-03 15:16:11 -07:00

11 lines
237 B
YAML

meta-llama/Llama-3.1-8B-Instruct:
- accuracy: 74.00
- spec_dec_algo: Eagle
accuracy: 74.00
- spec_dec_algo: NGram
accuracy: 74.00
deepseek-ai/DeepSeek-V3-Lite:
- accuracy: 77.00
- spec_dec_algo: MTP
accuracy: 77.00