[MTP] add test_draft_model_set_value_by_flags.py (#3741)

Author: Echo-Nie

tests/operators/test_draft_model_set_value_by_flags.py

@@ -0,0 +1,86 @@
+# 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.
+
+import unittest
+
+import numpy as np
+import paddle
+
+from fastdeploy.model_executor.ops.gpu import draft_model_set_value_by_flags
+
+
+class TestDraftModelSetValueByFlags(unittest.TestCase):
+    def setUp(self):
+        paddle.set_device("gpu")
+        np.random.seed(42)
+
+    def test_basic_update(self):
+        """
+        Test normal update behavior:
+        batch0 performs a decoder step, batch1 performs an encoder step
+        """
+        bs = 2
+        pre_id_length = 5
+        draft_tokens = paddle.to_tensor([[10, 11, 12], [20, 21, 22]], dtype="int64")
+        pre_ids_all = paddle.zeros([bs, pre_id_length], dtype="int64")
+        stop_flags = paddle.to_tensor([False, False], dtype="bool")
+        seq_lens_this_time = paddle.to_tensor([3, 1], dtype="int32")
+        seq_lens_encoder = paddle.to_tensor([0, 0], dtype="int32")
+        seq_lens_decoder = paddle.to_tensor([0, 0], dtype="int32")
+        step_idx = paddle.to_tensor([3, 1], dtype="int64")  # batch0 decoder, batch1 encoder
+
+        """ Call custom op """
+        draft_model_set_value_by_flags(
+            draft_tokens, pre_ids_all, stop_flags, seq_lens_this_time, seq_lens_encoder, seq_lens_decoder, step_idx
+        )
+
+        """
+        batch0: 3 tokens updated at decoder step
+        batch1: 1 token updated at encoder step
+        """
+        expected = np.array([[0, 10, 11, 12, 0], [0, 20, 0, 0, 0]], dtype=np.int64)
+
+        np.testing.assert_array_equal(pre_ids_all.numpy(), expected)
+        np.testing.assert_array_equal(seq_lens_this_time.numpy(), [1, 1])
+
+    def test_stop_flags(self):
+        """
+        batch0 is skipped (stop_flags=True), batch1 updates normally
+        """
+        bs = 2
+        pre_id_length = 4
+        draft_tokens = paddle.to_tensor([[5, 6], [7, 8]], dtype="int64")
+        pre_ids_all = paddle.zeros([bs, pre_id_length], dtype="int64")
+        stop_flags = paddle.to_tensor([True, False], dtype="bool")
+        seq_lens_this_time = paddle.to_tensor([2, 2], dtype="int32")
+        seq_lens_encoder = paddle.to_tensor([0, 0], dtype="int32")
+        seq_lens_decoder = paddle.to_tensor([0, 0], dtype="int32")
+        step_idx = paddle.to_tensor([1, 2], dtype="int64")
+
+        draft_model_set_value_by_flags(
+            draft_tokens, pre_ids_all, stop_flags, seq_lens_this_time, seq_lens_encoder, seq_lens_decoder, step_idx
+        )
+
+        """
+        batch0: no update due to stop flag
+        batch1: 2 tokens updated at decoder step
+        """
+        expected = np.array([[0, 0, 0, 0], [0, 7, 8, 0]], dtype=np.int64)
+
+        np.testing.assert_array_equal(pre_ids_all.numpy(), expected)
+        np.testing.assert_array_equal(seq_lens_this_time.numpy(), [2, 1])
+
+
+if __name__ == "__main__":
+    unittest.main()