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

Enable more unit tests for AMD CI. (#6983)

Browse Source
This commit is contained in:
Sai Enduri
2025-06-08 19:41:55 -07:00
committed by GitHub
parent 18efb5e8e0
commit 2c18642502
2 changed files with 56 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
test/srt/run_suite.py
View File

@@ -104,6 +104,29 @@ suites = {
TestFile("test_block_int8.py", 22),
TestFile("test_create_kvindices.py", 2),
TestFile("test_chunked_prefill.py", 313),
TestFile("test_embedding_openai_server.py", 141),
TestFile("test_eval_fp8_accuracy.py", 303),
TestFile("test_function_call_parser.py", 10),
TestFile("test_input_embeddings.py", 38),
TestFile("test_large_max_new_tokens.py", 41),
TestFile("test_metrics.py", 32),
TestFile("test_no_chunked_prefill.py", 108),
TestFile("test_no_overlap_scheduler.py", 234),
TestFile("test_penalty.py", 41),
TestFile("test_page_size.py", 60),
TestFile("test_pytorch_sampling_backend.py", 66),
TestFile("test_radix_attention.py", 105),
TestFile("test_reasoning_content.py", 89),
TestFile("test_enable_thinking.py", 70),
TestFile("test_request_length_validation.py", 31),
TestFile("test_retract_decode.py", 54),
TestFile("test_server_args.py", 1),
TestFile("test_skip_tokenizer_init.py", 117),
TestFile("test_torch_native_attention_backend.py", 123),
TestFile("test_triton_attention_backend.py", 150),
TestFile("test_update_weights_from_disk.py", 114),
TestFile("test_vertex_endpoint.py", 31),
TestFile("test_vision_chunked_prefill.py", 175),
],
"per-commit-2-gpu": [
TestFile("models/lora/test_lora_tp.py", 116),
@@ -116,13 +139,20 @@ suites = {
TestFile("test_verl_engine_2_gpu.py", 64),
],
"per-commit-2-gpu-amd": [
TestFile("models/lora/test_lora_tp.py", 116),
TestFile("test_data_parallelism.py", 73),
TestFile("test_mla_tp.py", 170),
TestFile("test_patch_torch.py", 19),
TestFile("test_update_weights_from_distributed.py", 103),
],
"per-commit-4-gpu": [
TestFile("test_local_attn.py", 250),
TestFile("test_pp_single_node.py", 150),
TestFile("test_verl_engine_4_gpu.py", 64),
],
"per-commit-4-gpu-amd": [
TestFile("test_pp_single_node.py", 150),
],
"per-commit-8-gpu": [
# Disabled deepep tests temporarily because it takes too much time.
# TODO: re-enable them after reducing the test time with compilation cache and smaller models.