diff --git a/.github/actions/commit_results/action.yaml b/.github/actions/commit_results/action.yaml index e5fa277b..00324f66 100644 --- a/.github/actions/commit_results/action.yaml +++ b/.github/actions/commit_results/action.yaml @@ -6,7 +6,7 @@ name: "Commit Test Results to a Branch" inputs: pretty_flags: required: false - default: "--metric='Latency (mean)' --metric='Bandwidth (mean)' --metric='GFLOP/s (mean)' -o pretty.md" + default: "--metric='Latency (mean)' --metric='Bandwidth (mean)' --metric='Throughput (mean)' -o pretty.md" input_csv: required: false default: "tests_latest.csv" diff --git a/iron/applications/llama_3.2_1b/test.py b/iron/applications/llama_3.2_1b/test.py index 18f4c602..8683615e 100644 --- a/iron/applications/llama_3.2_1b/test.py +++ b/iron/applications/llama_3.2_1b/test.py @@ -32,7 +32,7 @@ def generate_test_params(): @pytest.mark.metrics( TTFT=r"\[Prefill\]\s*Time to first token:\s*(?P[\d\.e\+-]+) s", - TPS=r"\[Decode\]\s*Tokens per second: (?P[\d\.e\+-]+)", + TPS=r"\[Decode\]\s*Tokens per second:\s*(?P[\d\.e\+-]+)", ) @pytest.mark.parametrize("prompt_len,num_tokens", params, ids=names) def test_llama_3_2_1b(prompt_len, num_tokens):