Jae-Won Chung
New leaderboard prototype
b10121d
raw
history blame contribute delete
326 Bytes
{
"Model": "codellama/CodeLlama-7b-hf",
"GPU": "NVIDIA A100-SXM4-40GB",
"TP": 2,
"PP": 1,
"Energy/req (J)": 24.174182842154483,
"Avg TPOT (s)": 0.08401189292034864,
"Token tput (tok/s)": 1366.373764737089,
"Avg Output Tokens": 102.46646341463415,
"Avg BS (reqs)": 127.20328638497652,
"Max BS (reqs)": 128
}