chore: import upstream snapshot with attribution

test/ai_edited_test/test_ai_decode_utils.py

@@ -0,0 +1,116 @@
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
+#
+# 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 test for paddle.nn.decode (BeamSearchDecoder, etc.)
+# Target: cover BeamSearchDecoder initialization and core methods
+
+import unittest
+
+import paddle
+from paddle import nn
+
+
+class TestBeamSearchDecoder(unittest.TestCase):
+    """Test BeamSearchDecoder initialization and basic methods.
+    BeamSearchDecoder is available as paddle.nn.BeamSearchDecoder.
+    Attributes are beam_size, start_token, end_token (no underscore prefix).
+    """
+
+    def setUp(self):
+        paddle.disable_static()
+
+    def test_beam_search_decoder_init(self):
+        """BeamSearchDecoder basic initialization."""
+        embed = nn.Embedding(100, 32)
+        cell = nn.SimpleRNNCell(32, 64)
+        output_layer = nn.Linear(64, 100)
+
+        decoder = nn.BeamSearchDecoder(
+            cell=cell,
+            start_token=1,
+            end_token=2,
+            beam_size=3,
+            embedding_fn=embed,
+            output_fn=output_layer,
+        )
+        self.assertIsNotNone(decoder)
+        self.assertEqual(decoder.beam_size, 3)
+        self.assertEqual(decoder.start_token, 1)
+        self.assertEqual(decoder.end_token, 2)
+
+    def test_beam_search_decoder_tile_beam(self):
+        """BeamSearchDecoder tile_beam_merge_with_batch static method."""
+        embed = nn.Embedding(100, 32)
+        cell = nn.SimpleRNNCell(32, 64)
+        output_layer = nn.Linear(64, 100)
+
+        decoder = nn.BeamSearchDecoder(
+            cell=cell,
+            start_token=1,
+            end_token=2,
+            beam_size=3,
+            embedding_fn=embed,
+            output_fn=output_layer,
+        )
+        x = paddle.randn([2, 5, 10], dtype='float32')
+        tiled = nn.BeamSearchDecoder.tile_beam_merge_with_batch(x, beam_size=3)
+        self.assertEqual(tiled.shape, [6, 5, 10])
+
+    def test_beam_search_decoder_no_embedding_fn(self):
+        """BeamSearchDecoder without embedding_fn."""
+        cell = nn.SimpleRNNCell(32, 64)
+        output_layer = nn.Linear(64, 100)
+
+        decoder = nn.BeamSearchDecoder(
+            cell=cell,
+            start_token=1,
+            end_token=2,
+            beam_size=2,
+            output_fn=output_layer,
+        )
+        self.assertIsNotNone(decoder)
+        self.assertEqual(decoder.beam_size, 2)
+
+
+class TestDynamicDecode(unittest.TestCase):
+    """Test dynamic_decode function.
+    dynamic_decode is available as paddle.nn.dynamic_decode.
+    """
+
+    def setUp(self):
+        paddle.disable_static()
+
+    def test_beam_search_decoder_with_dynamic_decode(self):
+        """Test that BeamSearchDecoder can be used with dynamic_decode."""
+        embed = nn.Embedding(100, 32)
+        cell = nn.SimpleRNNCell(32, 64)
+        output_layer = nn.Linear(64, 100)
+
+        decoder = nn.BeamSearchDecoder(
+            cell=cell,
+            start_token=1,
+            end_token=2,
+            beam_size=2,
+            embedding_fn=embed,
+            output_fn=output_layer,
+        )
+        # Verify decoder has the required interface for dynamic_decode
+        self.assertTrue(hasattr(decoder, 'step'))
+        self.assertTrue(hasattr(decoder, 'beam_size'))
+        self.assertTrue(hasattr(decoder, 'start_token'))
+        self.assertTrue(hasattr(decoder, 'end_token'))
+
+
+if __name__ == '__main__':
+    unittest.main()