test: add unit tests for Qwen3-VL mcore weight mapping

Add tests/unit/torch/export/test_mcore_qwen3vl_mapping.py covering
mapping types, lm_head root placement, language_model prefix presence,
layernorm replication annotations, TP sharding, and key-set symmetry.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Signed-off-by: Hung-Yueh Chiang <hungyuehc@nvidia.com>

Author: hychiang-git

tests/unit/torch/export/test_mcore_qwen3vl_mapping.py

@@ -0,0 +1,101 @@
+# SPDX-FileCopyrightText: Copyright (c) 2024-2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# 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.
+
+"""Unit tests for the Qwen3-VL mcore weight mapping (mcore_qwen3vl.py).
+
+Verifies that every key in qwen3vl_causal_lm_import / _export carries the
+``model.language_model.`` prefix (except ``lm_head.`` which stays at root),
+and that TP sharding and layernorm replication annotations are correct.
+"""
+
+import pytest
+
+from modelopt.torch.export.plugins.mcore_custom import (
+    COL_TP,
+    REPLICATE,
+    ROW_TP,
+    GatedMLPMerging,
+    GatedMLPSlicing,
+    NameRemapping,
+    QKVMerging,
+    QKVSlicing,
+)
+from modelopt.torch.export.plugins.mcore_qwen3vl import (
+    qwen3vl_causal_lm_export,
+    qwen3vl_causal_lm_import,
+)
+
+
+def test_mapping_types():
+    assert isinstance(qwen3vl_causal_lm_import["linear_qkv"], QKVMerging)
+    assert isinstance(qwen3vl_causal_lm_import["linear_fc1"], GatedMLPMerging)
+    assert isinstance(qwen3vl_causal_lm_export["linear_qkv"], QKVSlicing)
+    assert isinstance(qwen3vl_causal_lm_export["linear_fc1"], GatedMLPSlicing)
+
+
+def test_lm_head_at_root():
+    assert qwen3vl_causal_lm_import["output_layer"].target_name_or_prefix == "lm_head."
+    assert qwen3vl_causal_lm_export["output_layer"].target_name_or_prefix == "lm_head."
+
+
+@pytest.mark.parametrize(
+    "key",
+    [
+        "word_embeddings",
+        "final_layernorm",
+        "input_layernorm",
+        "linear_qkv",
+        "linear_proj",
+        "q_layernorm",
+        "k_layernorm",
+        "pre_mlp_layernorm",
+        "linear_fc1",
+        "linear_fc2",
+    ],
+)
+def test_language_model_prefix(key):
+    assert "model.language_model." in qwen3vl_causal_lm_import[key].target_name_or_prefix
+    assert "model.language_model." in qwen3vl_causal_lm_export[key].target_name_or_prefix
+
+
+@pytest.mark.parametrize(
+    "key",
+    ["input_layernorm", "q_layernorm", "k_layernorm", "pre_mlp_layernorm", "final_layernorm"],
+)
+def test_layernorm_replicated(key):
+    m = qwen3vl_causal_lm_import[key]
+    assert isinstance(m, NameRemapping)
+    assert m.func_kwargs == REPLICATE
+
+
+def test_tp_sharding():
+    assert qwen3vl_causal_lm_import["word_embeddings"].func_kwargs == COL_TP
+    assert qwen3vl_causal_lm_import["output_layer"].func_kwargs == COL_TP
+    assert qwen3vl_causal_lm_import["linear_proj"].func_kwargs == ROW_TP
+
+
+def test_export_no_parallel_config():
+    for key in [
+        "word_embeddings",
+        "final_layernorm",
+        "output_layer",
+        "input_layernorm",
+        "linear_proj",
+    ]:
+        assert "parallel_config" not in qwen3vl_causal_lm_export[key].func_kwargs
+
+
+def test_import_export_same_keys():
+    assert set(qwen3vl_causal_lm_import.keys()) == set(qwen3vl_causal_lm_export.keys())