Skip to content

Commit aefee25

Browse files
authored
Hide generate_time(ms) metric temporarily (#6314)
Due to pytorch/executorch#8576 (comment) As we cannot go back and update historical data, we could hide `generate_time(ms)` for a week or two till there are new data. Maybe it could also be hidden permanently if we decide to keep only TPS metric. ### Preview https://torchci-git-fork-huydhn-hide-generate-time-0f0c4f-fbopensource.vercel.app/benchmark/llms?repoName=pytorch%2Fexecutorch
1 parent 57cfb42 commit aefee25

File tree

2 files changed

+4
-1
lines changed

2 files changed

+4
-1
lines changed

torchci/components/benchmark/llms/SummaryPanel.tsx

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -133,7 +133,7 @@ export function SummaryPanel({
133133
];
134134

135135
const hasMode = data.length > 0 && "mode" in data[0] ? true : false;
136-
if (hasMode) {
136+
if (hasMode && benchmarkName === "TorchCache Benchmark") {
137137
columns.push({
138138
field: "mode",
139139
headerName: "Mode",

torchci/components/benchmark/llms/common.tsx

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -18,6 +18,9 @@ export const EXCLUDED_METRICS: string[] = [
1818
"cold_compile_time(s)",
1919
"warm_compile_time(s)",
2020
"speedup_pct",
21+
// TODO (huydhn): Hide generate_time(ms) metric temporarily because of
22+
// https://github.com/pytorch/executorch/issues/8576#issuecomment-2669706120
23+
"generate_time(ms)",
2124
];
2225
export const DEFAULT_MODEL_NAME = "All Models";
2326
export const SCALE = 2;

0 commit comments

Comments
 (0)