EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

Add accuracy and latency tests of eagle into CI (#3027)

Browse Source
This commit is contained in:
Lianmin Zheng
2025-01-21 02:55:14 -08:00
committed by GitHub
parent ec1c21cdc4
commit a4331cd260
7 changed files with 186 additions and 123 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
.github/workflows/pr-test.yml vendored
View File

@@ -128,7 +128,7 @@ jobs:
timeout-minutes: 10
run: |
cd test/srt
python3 -m unittest test_bench_one_batch.TestBenchOneBatch.test_default
python3 -m unittest test_bench_one_batch.TestBenchOneBatch.test_bs1
- name: Benchmark online latency
timeout-minutes: 10
@@ -148,6 +148,13 @@ jobs:
cd test/srt
python3 -m unittest test_bench_serving.TestBenchServing.test_offline_throughput_non_stream_small_batch_size
- name: Benchmark online latency (EAGLE)
timeout-minutes: 10
run: |
cd test/srt
python3 -m unittest test_bench_serving.TestBenchServing.test_online_latency_eagle
performance-test-1-gpu-part-2:
if: github.repository == 'sgl-project/sglang' || github.event_name == 'pull_request'
runs-on: 1-gpu-runner
@@ -196,7 +203,13 @@ jobs:
timeout-minutes: 10
run: |
cd test/srt
python3 -m unittest test_bench_one_batch.TestBenchOneBatch.test_moe_default
python3 -m unittest test_bench_one_batch.TestBenchOneBatch.test_moe_tp2_bs1
- name: Benchmark single latency + torch.compile (TP=2)
timeout-minutes: 10
run: |
cd test/srt
python3 -m unittest test_bench_one_batch.TestBenchOneBatch.test_torch_compile_tp2_bs1
- name: Benchmark offline throughput (TP=2)
timeout-minutes: 10
@@ -210,6 +223,7 @@ jobs:
cd test/srt
python3 -m unittest test_bench_serving.TestBenchServing.test_moe_offline_throughput_without_radix_cache
accuracy-test-1-gpu:
if: github.repository == 'sgl-project/sglang' || github.event_name == 'pull_request'
runs-on: 1-gpu-runner