init v0.23.0

Signed-off-by: Sun Ruoxi <sunruoxi@4paradigm.com>

tests/e2e/pull_request/four_card/_310p/test_dense_model_310p.py

@@ -0,0 +1,49 @@
+#
+# Copyright (c) 2026 Huawei Technologies Co., Ltd. All Rights Reserved.
+# Copyright 2023 The vLLM team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# This file is a part of the vllm-ascend project.
+
+from tests.e2e.conftest import VllmRunner
+
+
+def test_qwen3_dense_tp2_fp16():
+    example_prompts = [
+        "Hello, my name is",
+    ]
+    max_tokens = 5
+    with VllmRunner(
+        "Qwen/Qwen3-8B",
+        tensor_parallel_size=2,
+        enforce_eager=True,
+        dtype="float16",
+        max_model_len=16384,
+    ) as vllm_model:
+        vllm_model.generate_greedy(example_prompts, max_tokens)
+
+
+def test_qwen3_dense_tp4_w8a8():
+    example_prompts = [
+        "Hello, my name is",
+    ]
+    max_tokens = 5
+    with VllmRunner(
+        "vllm-ascend/Qwen3-32B-W8A8",
+        tensor_parallel_size=4,
+        enforce_eager=True,
+        dtype="float16",
+        quantization="ascend",
+        max_model_len=16384,
+    ) as vllm_model:
+        vllm_model.generate_greedy(example_prompts, max_tokens)

tests/e2e/pull_request/four_card/_310p/test_moe_model_310p.py

@@ -0,0 +1,74 @@
+#
+# Copyright (c) 2026 Huawei Technologies Co., Ltd. All Rights Reserved.
+# Copyright 2023 The vLLM team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# This file is a part of the vllm-ascend project.
+
+
+import pytest
+
+from tests.e2e.conftest import VllmRunner
+
+
+def test_qwen3_moe_tp4_fp16():
+    example_prompts = [
+        "Hello, my name is",
+    ]
+    max_tokens = 5
+    with VllmRunner(
+        "Qwen/Qwen3-30B-A3B",
+        tensor_parallel_size=4,
+        enforce_eager=True,
+        dtype="float16",
+        max_model_len=16384,
+        max_num_batched_tokens=2048,
+        max_num_seqs=256,
+    ) as vllm_model:
+        vllm_model.generate_greedy(example_prompts, max_tokens)
+
+
+def test_qwen3_moe_tp2_w8a8():
+    example_prompts = [
+        "Hello, my name is",
+    ]
+    max_tokens = 5
+    with VllmRunner(
+        "vllm-ascend/Qwen3-30B-A3B-W8A8",
+        tensor_parallel_size=2,
+        enforce_eager=True,
+        dtype="float16",
+        quantization="ascend",
+        max_model_len=16384,
+        max_num_batched_tokens=2048,
+        max_num_seqs=256,
+    ) as vllm_model:
+        vllm_model.generate_greedy(example_prompts, max_tokens)
+
+
+@pytest.mark.skip("Probabilistic failure, need fix")
+def test_qwen3_5_moe_tp4_fp16():
+    example_prompts = [
+        "Hello, my name is",
+    ]
+    max_tokens = 5
+    with VllmRunner(
+        "Qwen/Qwen3.5-35B-A3B",
+        tensor_parallel_size=4,
+        enforce_eager=True,
+        dtype="float16",
+        max_model_len=16384,
+        max_num_batched_tokens=2048,
+        max_num_seqs=256,
+    ) as vllm_model:
+        vllm_model.generate_greedy(example_prompts, max_tokens)

tests/e2e/pull_request/four_card/_310p/test_vl_model_310p.py

@@ -0,0 +1,34 @@
+#
+# Copyright (c) 2026 Huawei Technologies Co., Ltd. All Rights Reserved.
+# Copyright 2023 The vLLM team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# This file is a part of the vllm-ascend project.
+
+import os
+import sys
+
+from tests.e2e.conftest import wait_until_npu_memory_free
+
+current_dir = os.path.dirname(os.path.abspath(__file__))
+full_dir = os.path.dirname(os.path.dirname(current_dir))
+sys.path.insert(0, full_dir)
+
+# ruff: noqa: E402
+from tests.e2e.pull_request.utils_310p import run_vl_model_test
+
+
+@wait_until_npu_memory_free(target_free_percentage=0.7)
+def test_qwen3_vl_8b_tp2_fp16():
+    """Qwen3-VL-8B dual-card FP16 test"""
+    run_vl_model_test(model_name="Qwen/Qwen3-VL-8B-Instruct", tensor_parallel_size=2, max_tokens=5)