ci: migrate 1-GPU model tests to test/registered/models/ (#16414)

2026-01-04 18:08:01 -08:00
parent 249c356331
commit f8411ded6e
17 changed files with 111 additions and 24 deletions
@@ -1,3 +1,9 @@
+from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
+
+# Model tests for compressed tensors (FP8)
+register_cuda_ci(est_time=42, suite="stage-b-test-small-1-gpu")
+register_amd_ci(est_time=42, suite="stage-b-test-small-1-gpu")
+
 import unittest
 from types import SimpleNamespace

@@ -1,3 +1,9 @@
+from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
+
+# Cross encoder model tests
+register_cuda_ci(est_time=100, suite="stage-b-test-small-1-gpu")
+register_amd_ci(est_time=150, suite="stage-b-test-small-1-gpu")
+
 import multiprocessing as mp
 import random
 import unittest
@@ -1,3 +1,13 @@
+from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
+
+# Embedding model tests
+register_cuda_ci(est_time=73, suite="stage-b-test-small-1-gpu")
+register_amd_ci(
+    est_time=73,
+    suite="stage-b-test-small-1-gpu",
+    disabled="see https://github.com/sgl-project/sglang/issues/11127",
+)
+
 # Copyright 2023-2024 SGLang Team
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -1,3 +1,8 @@
+from sglang.test.ci.ci_register import register_cuda_ci
+
+# Encoder embedding model tests (CUDA only)
+register_cuda_ci(est_time=221, suite="stage-b-test-small-1-gpu")
+
 # Copyright 2023-2024 SGLang Team
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -1,3 +1,8 @@
+from sglang.test.ci.ci_register import register_cuda_ci
+
+# Generation model tests (CUDA only)
+register_cuda_ci(est_time=103, suite="stage-b-test-small-1-gpu")
+
 # Copyright 2023-2024 SGLang Team
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -1,3 +1,9 @@
+from sglang.test.ci.ci_register import register_cuda_ci
+
+# NVIDIA Nemotron Nano V2 VL model tests (CUDA only)
+# GSM8k + MMMU evaluation
+register_cuda_ci(est_time=214, suite="stage-b-test-small-1-gpu")
+
 import unittest

 from sglang.test.kits.gsm8k_accuracy_kit import GSM8KMixin
@@ -1,3 +1,9 @@
+from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
+
+# Qwen model tests
+register_cuda_ci(est_time=90, suite="stage-b-test-small-1-gpu")
+register_amd_ci(est_time=82, suite="stage-b-test-small-1-gpu")
+
 import unittest
 from types import SimpleNamespace

@@ -1,3 +1,9 @@
+from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
+
+# Reward model tests
+register_cuda_ci(est_time=103, suite="stage-b-test-small-1-gpu")
+register_amd_ci(est_time=132, suite="stage-b-test-small-1-gpu")
+
 # Copyright 2023-2024 SGLang Team
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -1,3 +1,9 @@
+from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
+
+# Transformers fallback model tests
+register_cuda_ci(est_time=245, suite="stage-b-test-small-1-gpu")
+register_amd_ci(est_time=320, suite="stage-b-test-small-1-gpu")
+
 import dataclasses
 import multiprocessing as mp
 import unittest
@@ -1,3 +1,9 @@
+from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
+
+# VLM (Vision Language Model) tests
+register_cuda_ci(est_time=270, suite="stage-b-test-small-1-gpu")
+register_amd_ci(est_time=387, suite="stage-b-test-small-1-gpu")
+
 import argparse
 import random
 import sys
@@ -18,7 +18,7 @@ HW_MAPPING = {
 # Per-commit test suites (run on every PR)
 PER_COMMIT_SUITES = {
    HWBackend.CPU: ["default"],
-    HWBackend.AMD: ["stage-a-test-1"],
+    HWBackend.AMD: ["stage-a-test-1", "stage-b-test-small-1-gpu"],
    HWBackend.CUDA: [
        "stage-a-test-1",
        "stage-b-test-small-1-gpu",
@@ -15,16 +15,6 @@ suites = {
        TestFile("layers/attention/mamba/test_causal_conv1d.py", 25),
        TestFile("layers/attention/mamba/test_mamba_ssm.py", 7),
        TestFile("layers/attention/mamba/test_mamba_ssm_ssd.py", 13),
-        TestFile("models/test_compressed_tensors_models.py", 42),
-        TestFile("models/test_cross_encoder_models.py", 100),
-        TestFile("models/test_embedding_models.py", 73),
-        TestFile("models/test_encoder_embedding_models.py", 221),
-        TestFile("models/test_generation_models.py", 103),
-        TestFile("models/test_nvidia_nemotron_nano_v2_vl.py", 214),  # GSM8k + MMMU
-        TestFile("models/test_qwen_models.py", 90),
-        TestFile("models/test_reward_models.py", 103),
-        TestFile("models/test_transformers_models.py", 245),
-        TestFile("models/test_vlm_models.py", 270),
        TestFile("openai_server/basic/test_openai_embedding.py", 70),
        TestFile("openai_server/basic/test_openai_server.py", 184),
        TestFile("openai_server/basic/test_protocol.py", 3),
@@ -195,12 +185,6 @@ suite_amd = {
        # TestFile("lora/test_lora_cuda_graph.py", 250), # Disabled temporarily, see https://github.com/sgl-project/sglang/issues/13107
        # TestFile("lora/test_lora_qwen3.py", 97), # Disabled temporarily, see https://github.com/sgl-project/sglang/issues/13107
        TestFile("dllm/test_llada2_mini_amd.py", 520),
-        TestFile("models/test_compressed_tensors_models.py", 42),
-        TestFile("models/test_cross_encoder_models.py", 150),
-        TestFile("models/test_qwen_models.py", 82),
-        TestFile("models/test_reward_models.py", 132),
-        TestFile("models/test_transformers_models.py", 320),
-        TestFile("models/test_vlm_models.py", 387),
        TestFile("openai_server/basic/test_openai_embedding.py", 141),
        TestFile("openai_server/basic/test_openai_server.py", 149),
        TestFile("openai_server/basic/test_protocol.py", 10),
@@ -261,7 +245,6 @@ suite_amd = {
        TestFile("test_video_utils.py", 8),
        # Disabled temporarily
        # TestFile("test_vlm_input_format.py", 300),
-        # TestFile("models/test_embedding_models.py", 73), # Disabled temporarily, see https://github.com/sgl-project/sglang/issues/11127
        # TestFile("openai_server/features/test_openai_server_hidden_states.py", 240),
        # TestFile("rl/test_update_weights_from_tensor.py", 48),
        # TestFile("test_no_overlap_scheduler.py", 234), # Disabled temporarily and track in #7703