You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I encountered an issue while testing eagle-1 and llama-2-7b on the 4090 using the mt_beach dataset. The speed ratio is normal, but the alpha values do not match those reported in the research paper.
I encountered an issue while testing eagle-1 and llama-2-7b on the 4090 using the mt_beach dataset. The speed ratio is normal, but the alpha values do not match those reported in the research paper.
python -m eagle.evaluation.gen_ea_alpha_llama2chat --ea-model-path /models/EAGLE-llama2-chat-7B --base-model-path models/modelscope/Llama-2-7b-chat-ms --model-id ea-alpha-llama-2-chat-7b-fp16 --temperature 0.0
python mt_bench/alpha.py
The text was updated successfully, but these errors were encountered: